Control of Dioxins (and otherOrganochlorines) from the Pulp and PaperIndustry under the Clean Water Act andLead in Soil at Superfund Mining Sites: TwoCase Studies in EPA's Use of Science
Mark R. Powell
Discussion Paper 97-08
March 1997 (Revised)
1616 P Street, NWWashington, DC 20036Telephone 202-328-5000Fax 202-939-3460
Discussion papers are research materials circulated bytheir authors for purposes of information and discussion.They have not undergone formal peer review or theeditorial treatment accorded RFF books and otherpublications.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites i
Abstract
This paper discusses EPA’s acquisition and use of science in addressing dioxins (and otherorganochlorines) from the pulp and paper industry under the Clean Water Act and lead in soil atlarge Superfund mining sites. The common thread between both cases is the challenge posed byadministering national pollution control programs while considering site-by-site variability infactors that influence environmental risks. In the first case study, high levels of dioxin in fishdownstream of pulp and paper mills were inadvertently detected in 1983 as part of an EPA effortto determine background levels of dioxin in areas presumed to be relatively uncontaminated. Thesefindings quickly got the release of dioxins from pulp and paper mills on EPA’s research agenda.News reports beginning in 1987 elevated the issue onto the regulatory agenda, but more than adecade has passed without EPA taking final regulatory action. Meanwhile, the pulp and paperindustry has dramatically reduced, but not eliminated, dioxin discharges from mills. The keyscientific issue now confronting EPA decisionmakers is how much weight to give to a water qualityindicator called AOX. AOX is not statistically related to dioxin at the levels under consideration.Environmentalists justify using AOX because it serves as a surrogate measure for the entiretoxicologically uncharacterized “soup” of organocholorines discharged from bleaching mills.Additionally, EPA estimates that discharges of dioxin from plants at levels below the analyticaldetection limits will continue to result in exceedances of stringent federal ambient water qualitycriteria under some local conditions. Industry counters that reductions in AOX do not achieve anymeasurable or monetizable environmental benefits. This case illustrates EPA’s use of science toevaluate the cost-effectiveness of nominally technology-based water pollution controls. In thesecond case study, the Superfund program does not have the option of following its standardoperating procedures for evaluating risks and determining Preliminary Remediation Goals for lead-contaminated sites because EPA has no numerical health-based standard for ingested lead (theagency’s goal for lead is based on the level of lead in children’s bloodstream). The study,therefore, illuminates the challenges and opportunities posed by developing and using rigorous site-specific scientific information. Potentially Responsible Parties (PRPs) generated rodent bioassaydata which suggested that the bioavailability of lead in soil at mining sites would be much lowerthan EPA’s default assumption. However, the agency disputed the validity of using mature rodentsas animals models for the population of concern, children. In response, EPA conductedexperiments with juvenile swine. The results indicated considerable variability in thebioavailability of lead in soil among the sites tested, with some higher, some lower, and some aboutthe same as the agency’s default assumption. Consequently, EPA cannot generalize across siteswhere similar mining activities occurred or draw any general distinctions between different types ofmining sites, as had been presumed. This case illustrates that selection of the most appropriateanimal model for toxicological studies involves tradeoffs between cost, experimental power andcontrol, fidelity to human physiology, and the value of information for decisionmaking.Determination of the “optimal” animal model depends on the evaluative criterion being used.Although the new scientific data generated by EPA suggests higher bioavailability of lead in soil atsome sites than the agency’s default assumption, in terms of the final remedy selection, it appearsthat all of the results will be either beneficial or essentially neutral to Large Area Lead Site PRPsbecause EPA deems the cost of removing the contaminated soil to be excessive.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites ii
Abstract
Table of Contents
Introduction ........................................................................................................................ iii
A. Control of Dioxins (and Other Organochlorines) from the Pulp and Paper IndustryUnder the Clean Water Act ..........................................................................................1
List of Abbreviations............................................................................................59
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites iii
INTRODUCTION
The case studies included in this discussion paper are part of a project thatResources for the Future (RFF) is conducting under a cooperative agreement with theU.S. Environmental Protection Agency (EPA) and with general support from RFF. Thecase studies were originally vetted as RFF Discussion Paper 97-08 in 1996, and thisrevised version of the discussion paper reflects many useful comments and correctionssupplied by reviewers.
The overall study is broadly concerned with the acquisition and use of scientificinformation by the Environmental Protection Agency in regulatory decisionmaking. Theoverall study focuses chiefly on national rulemaking (e.g., setting National Ambient AirQuality Standards and banning pesticides or toxic substances), as opposed to site-specificdecisionmaking (e.g., Superfund remedy selection). For the purposes of this study,environmental “science” refers to information that can be used in assessing risks to humanhealth, welfare, and the environment. (Therefore, economic and engineering informationare not a chief focus of this study.) The project aims to help policymakers and othersbetter understand the factors and processes that influence EPA's acquisition and use ofscience in national rulemaking so that they can better evaluate recommendations forimproving environmental regulatory institutions, policies, and practices.
In all, eight case studies will be included as appendices to the full report:
• 1987 Revision of the National Ambient Air Quality Standard forParticulates (NAAQS)
• 1993 Decision Not to Revise the NAAQS for Ozone• 1991 Lead/Copper Rule under the Safe Drinking Water Act (SDWA)• 1995 Decision to Pursue Additional Research Prior to Revising the Arsenic
Standard under SDWA• 1983/4 Suspensions of Ethylene Dibromide under the Federal Insecticide,
Fungicide, and Rodenticide Act• 1989 Asbestos Ban & Phaseout Rule under the Toxic Substances Control
Act• Control of Dioxins (and other Organochlorines) from Pulp & Paper
effluents under the Clean Water Act (as part of the combined air/water“cluster rule” proposed in 1993)
• Lead in Soil at Superfund Mining Sites
The case studies were selected in consultation with informal advisors to the projectand are not intended as a random or representative sample of EPA regulatory decisions.None of the case studies could be fairly characterized as routine or pedestrian. As agroup, the cases tend toward the “high-profile” end of the distribution of EPA decisions.Nevertheless, among the case studies, there is some variability in the political andeconomic stakes involved and in the level of development of the underlying science. Thecases selected involve each of the “national” environmental regulatory statutes (Clean Air
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites iv
Act; Safe Drinking Water Act; Toxic Substances Control Act; Federal Insecticide,Fungicide, and Rodenticide Act; and Clean Water Act), and two cases involve decisions tomaintain the status quo (ozone and arsenic), as opposed to the remainder of the caseswhich involve decisions to change from the status quo.
Methodology
Development of the case studies was based on literature review and interviewswith persons inside and outside EPA. The number of interviewees per case study variedroughly from a half dozen to a dozen. There was an effort to ensure balance in the groupof respondents for any particular case study, but because of the relatively small number ofrespondents and the non-random nature of the selection process, extreme caution shouldbe taken in interpreting the numerical response summaries that are reported. Interviewswere conducted primarily using a structured questionnaire format, but in some cases,comments were sought from specific individuals regarding particular issues instead of thecase as a whole. In addition to interviews specific to particular case studies, interviewswere also conducted for the overall study to elicit the views of current and formerpolicymakers, senior scientists, specialists in regulatory science issues, and othersregarding EPA’s acquisition and use of science. The case studies also incorporate manycomments and insights from these interviewees.
In all instances, interviewees were given the option of speaking for attribution oroff-the-record, and almost all respondents elected to speak off-the-record. A completelisting of the more than 100 interviewees for the overall study will be included as anappendix to the final report. The selection of interviewees considered that individualsfrom the bench scientist through the agency staff analyst to the politically appointeddecisionmaker, as well as advocates from outside the agency, would provide informativeperspectives. Among the wide range of interviewees were: 5 of 6 former EPAAdministrators, 4 current or former Deputy Administrators, and 5 current or formerAssistant Administrators; 4 current or former congressional staff; several current andformer EPA Science Advisory Board members; various representatives of industry andenvironmental advocacy groups; environmental journalists; and academics from thediverse fields of biology, public health, economics, political science, psychology, andphilosophy. But to better understand the processes occurring within the agency,interviewees were disproportionately selected from among current and former EPAofficials.
A prominent feature of the case studies consists of an effort to map the origins,flow, and effect of scientific information relating to a particular decision. To accomplishthis, the case studies make use of an extended analogy to fate and transport modeling. Asused in risk assessment, this modeling procedure predicts the movement andtransformation of pollutants from their point of origin to their ultimate destination. Thus,to extend the analogy, one can imagine universities and research institutes “emitting”scientific findings, which are disseminated and “transformed” by the media and consultantsoutside the agency. (An alternative pattern is when scientific findings are generated within
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites v
EPA by agency scientists.) Science can enter EPA through multiple “exposure routes,”which assimilate information differently; once inside the agency, information is“metabolized” prior to its “delivery” to the “target organ” (the decision-maker). This fateand transport terminology is adopted because it is part of the vernacular of many of thoseproviding the information and of many of the ultimate users of the study results. Figure Apresents a simplified model of the fate and transport of science in environmental regulationfor illustrative purposes.
Figure A. Fate and Transport of Science in Environmental Regulation
Making use of these conceptual models, we attempt to address questionsspecifically about the scientific information in each of the case studies, such as: what arethe sources and their relative contributions? where are the points-of-entry? who are thegatekeepers? what is the internal transport mechanism? how is the informationtransformed as it flows through the agency? what does and doesn’t get communicated tothe decisionmaker? and where and how is the information ultimately applied?
Comments on the case studies should be addressed to:
Mark Powell, FellowCenter for Risk ManagementResources for the Future1616 P St., NWWash., DC 20036tel: 202/328-5070fax: 202/939-3460email: [email protected]
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 1
A. Control of Dioxins (and Other Organochlorines) from the Pulp andPaper Industry under the Clean Water Act
1. Background
The term dioxin encompasses a family of organic chemical compounds known asdibenzo-p-dioxins. The dioxins of greatest environmental and public health concern arehalogenated dioxins.1 Because they are the most common, most attention is focused onthe group of 75 chlorinated dioxins. Dioxins are not deliberately manufactured, but are abyproduct of combustion, some chemical manufacturing, some bleaching of pulp andpaper, and other industrial processes involving chlorine and other halogens. In the U.S.,municipal and medical waste incineration are the dominant known sources of dioxin (EPA1994a), but the total releases of dioxin from all sources (including natural sources such asforest fires) is highly uncertain. Dioxin became notorious in the 1970’s when it wasidentified in the U.S. as “the most potent animal carcinogen ever tested.” As one observerphrased it, dioxin earned the reputation as the “Darth Vader” of chemicals (Roberts 1991).In the 1980s, however, Canada and European countries set dioxin limits less stringent thanEPA’s by two or three orders of magnitude. Officials in these countries concluded that adifferent cancer model applied to dioxin. More recently, attention has focused on theenvironmental and non-cancer effects of dioxin and dioxin-like substances that may mimichormones and act as “endocrine disruptors.”
Dioxin discharges into surface waters from pulp and paper mills aroseunexpectedly as a regulatory issue more than a decade ago. In 1982, EPA promulgatedClean Water Act (CWA) effluent limitations and technology-based standards (“effluentguidelines”) for most of the pulp, paper, and paperboard industry.2 A year later, as partof the EPA’s overall “Dioxin Strategy,” the agency initiated a national survey ofenvironmental dioxin levels. In the process of testing what were believed to be “referencestreams” to determine background dioxin concentrations in fish in relativelyuncontaminated waters, the agency detected surprisingly high levels of dioxin.3 Accordingto an EPA official, the reference streams where fish had elevated dioxin concentrationshad one feature in common, “when you looked upstream, they all had chlorine bleaching
Over time, the list of toxic water pollutants of concern related to chlorine pulpbleaching was broadened to include a variety of more abundant chlorinated organiccompounds (organochlorines). These include polychlorinated phenolic compounds, whichare considered representative of a various polychlorinated organic materials that mayaccumulate in food chains, and chloroform, a volatile organic compound. Indicative of the
1 Halogens include chlorine, bromine, iodine, etc.2 Effluents are wastewater discharges into surface waters.3 For example, dioxin concentrations in fish in a Wisconsin reservoir were more than 50 ppt (parts pertrillion), leading the state to close a commercial fishery. Samples in Maine and Minnesota found dioxinconcentration in fish of up to 85 ppt (Harrison and Hoberg 1991). By comparison, measured backgroundlevels of dioxin in fish are 0-2 ppt (EPA 1994a).
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 2
relative magnitude of their production by the U.S. pulp and paper industry, discharges ofdioxins and a group of dioxin-like chemicals called furans are measured in terms of gramsper year, while the discharges of other organochlorines are expressed in units of metrictons per year. Despite this disparity in the quantity of environmental releases, dioxins andfurans have dominated the debate over regulatory controls of the effluents from pulp andpaper plants that use chlorine bleaching because chlorinated phenols and volatileorganochlorines are estimated to be very much less toxic. Some individuals and groupsremained concerned, however, about the heterogeneous soup of organochlorinesdischarged in bulk from pulp and paper mills because most of these compounds have notbeen toxicologically analyzed and because the chemical transformations organochlorinesundergo in the environment are not fully understood. Staking out a precautionary positionin the face of scientific uncertainty, some interested and affected parties argue that allorganochlorines should be considered “guilty until proven innocent.”
The regulatory control of dioxins, furans, and other organochlorines dischargedfrom pulp and paper mills into surface waters traces its origins back to October 1984,when the Environmental Defense Fund (EDF) and the National Wildlife Federation(NWF) filed a citizen’s petition under the Toxic Substances Control Act (TSCA, Sec. 21).The petition requested that EPA regulate dioxins and furans from all known sources.4 (Atthe time, despite the questions raised by the detection of dioxin in streams below pulp andpaper mills, the bleaching plants were not yet recognized as a source of dioxins andfurans.) EPA denied the petition, prompting a 1985 lawsuit by EDF and NWF (EDF v.Thomas, DC Dist. Court, Civ. No. 85-0973). Following a series of news reports aboutEPA’s cooperation with industry to investigate the formation and release of dioxins atpulp and paper plants and a 1987 front page story in the New York Times regarding thedetection of dioxin in household paper products, EPA signed a consent decree with theplaintiffs in 1988. The agreement required EPA to perform a comprehensive riskassessment of dioxins and furans considering sludges, water effluent, and products madefrom pulp produced at 104 bleaching pulp mills. The agreement also required the agencyto propose regulations under TSCA (Sec. 6) to control pulp sludge disposal and under theClean Water Act to address discharges of dioxins and furans into surface waters from themills by October 31, 1993 (as amended in 1992). The agency’s 1993 proposal to controldioxin and furan releases into surface waters is the primary focus of this case study. Theproposal was submitted as a combined set of water effluent limitations and standards andnational emission standards for hazardous air pollutants for the pulp, paper, andpaperboard industrial sector (also called the proposed “pulp and paper cluster rule,” Fed.Reg., Vol. 58, pp. 66078-66216). The pulp and paper cluster rule had not been finalizedas of press time. But it appears that the crucial subplot for the effluent limits involves anarcane debate over a Swedish water quality test measure called AOX.
4 The term furans refers to chlorinated dibenzofurans.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 3
Regulation of Toxic Water Pollutants
The goal of the CWA (also known as the 1972 Federal Water Pollution ControlAct (FWPCA) Amendments) is to eliminate entirely discharges of pollutants from pointsources (i.e., individual discharging facilities) into surface waters. Although eliminatingpollutant discharges may be achievable under some circumstances through processchanges that prevent pollutant formation or recycle wastes, the goal is largely rhetorical.The statutory goal of eliminating discharges has potential distorting effects on the use ofscience because achieving the goal does not require point sources to eliminate alldischarges into surface waters. Consequently, attainment depends to some extent on whatsubstances are classified as pollutants subject to regulation under the statute.Furthermore, pursuing discharge elimination from one point source may result in offsettingreleases of pollutants. For example, on-site waste recovery to prevent surface waterdischarges may require extra energy inputs, resulting in additional releases of contaminantsto the atmosphere.
The CWA contains both “water-quality based” regulatory controls, which varyaccording to the designated use (e.g., drinking water source, fishable, swimmable) andattributes (e.g., volume and rate of flow) of the receiving water body, and “technology-based” effluent standards that are achievable using available pollution control technology.Legally, the environmental quality standards dominate the technology-based standards inthe sense that additional regulatory action may be required if the technology-based limitsdo not achieve the ambient quality standard in a specific location. In practice, thetechnology-based standards are emphasized. This is due in part to the practical difficultiesexperienced prior to 1972 with state attempts to control surface water pollution. TheFWPCA relied on water quality standards which required state regulatory authorities todemonstrate that a given level of pollution was “unreasonable” or “unacceptable” underlocal environmental and socioeconomic conditions. Under the CWA, Congress hasemphasized the approach of the technology-based effluent standards that “do not quibblewith judgments of reasonableness” (Fogarty 1991). The emphasis on technology-basedstandards also avoids the potentially greater time and cost associated with developing,administering, and complying with myriad geographically-specific pollutant dischargelimits that must be tailored to meet ambient water quality standards.
Under the 1972 provisions, EPA was to develop a list of national standards fortoxic water pollutants that would be applied without regard to industrial source.Implementation of this chemical-by-chemical approach was more difficult than Congressexpected, and dissatisfaction with the progress lead to litigation and, eventually, a 1976consent decree between the Natural Resources Defense Council and EPA.5 The approachlaid out in this settlement was ratified in the 1977 CWA Amendments. Sec. 307 of theCWA now requires Best Available Technology (BAT) economically achievable byindustrial sector to limit toxic pollutant effluents from point sources into surface waters.The settlement originally identified a list of 65 “toxic” chemicals and classes of chemicals,
5 NRDC et al. v. Train, 8 ERC 2120 (D.D.C. 1976). Later modified as 12 ERC 1833 (D.D.C. 1979).
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 4
which were later subdivided into 129 individual substances or “priority” pollutants (CRS1993).6 Dioxin (TCDD - 2,3,7,8-tetrachlorodibenzo-p-dioxin) was originally placed onboth lists of toxic pollutants.7
The CWA directs EPA to develop BAT for toxic water pollutants “that will resultin reasonable further progress toward the national goal of eliminating discharges” (Sec.301(b)(2)). Factors to be considered in developing BAT for toxic water pollutants includethe affordability of achieving effluent reductions (“economic achievability”), engineeringcriteria, non-water quality environmental impacts, and “such other factors as theAdministrator deems appropriate” (Sec. 304(b)(2)). The BAT basis for regulating toxicpollutants is in contrast to the control of “conventional” pollutants (e.g., suspended solidsand fecal coliform). Under Sec. 304 of the CWA, conventional pollutant limits areachieved by Best Conventional Pollutant Control Technology (BCT). Determination ofBCT depends on the relationship between costs and benefits (essentially a BAT standardmoderated by a test of economic reasonableness) (Fogarty 1991). Thus BAT control oftoxic pollutants is intended to be less sensitive to cost considerations than BCT, but itacknowledges that alternative technologies can be compared in terms of environmentalbenefits. (That is, for one technology to be the “best” it must achieve environmentalbenefits superior to another technology.) Section 307(a) also allows EPA to impose morestringent toxic effluent standards if the BAT standard is inadequate to protect humanhealth with an “ample margin of safety.” For some toxic pollutants, however, the onlymeans of providing any margin of safety (ample or otherwise) may be to prohibitdischarges altogether because there may be no discernible threshold level of incrementalexposure below which no adverse effects will occur.8
Through its 1993 proposed pulp and paper effluent regulations, EPA sought tolimit the precursors to the formation of dioxins, furans, and other organochlorines in thepulp and paper manufacturing process. The technology-based approach proposed by theagency involves: 1) substituting elemental chlorine with chlorine dioxin or other bleachingagents (e.g., peroxide or ozone) and 2) reducing the extent of chlorine bleaching requiredto achieve a given quality of product through alternative means of pulp delignification(i.e., extended cooking or oxygen delignification prior to chlorine bleaching). The agencyestimates that its proposed effluent limits for the pulp and paper industry would reduce,
6See 40 CFR 401.15 and Fed. Reg. Vol. 57, pp. 60911-15 respectively for complete lists. EPA has sincereduced the number of priority pollutants to 126. Priority pollutants are carcinogens, suspectedcarcinogens, or pollutants known to be seriously toxic at low levels. The priority pollutant list originatedfrom a 1975 EPA water toxics regulatory strategy developed in response to the NRDC lawsuit (CRS1993).7 Although furan (TCDF - 2,3,7,8-tetrachlorodibenzofuran) is not explicitly listed as a priority pollutant,EPA treats it as a dioxin-like compound. The International Joint Commission (IJC) has identified TCDDand TCDF as two of eleven “Critical Pollutants” for the Great Lakes (AET 1995).8 Toxic effluent standards are pollutant-specific, nationally uniform, and applicable across all categories ofindustry and all dischargers. By 1976, EPA had promulgated such standards for aldrin/dieldrin, DDT,endrin, toxaphene, benzidine, and polychlorinated biphenyls (PCBs), but stringent procedural andscientific requirements have prevented more extensive development of toxic effluent standards under Sec.307(a)(2) (Fogarty 1991).
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 5
but not eliminate exceedances of health-based water standards for dioxins and furans (seeTable A-4 below). However, EPA’s ambient water quality criteria are not necessarily thelast word. Under the Clean Water Act, EPA and the States share responsibility andauthority for: setting risk-based ambient water quality standards; identifying specificsegments of water bodies where technology-based pollutant controls may be inadequate toachieve uses designated by the States; and developing strategies for achieving ambientwater quality standards in these impaired waters.
Section 303 of the 1987 Clean Water Act Amendments required States to adoptbinding numeric criteria for all priority pollutants in cases where discharges couldreasonably be expected to interfere with the designated use of water bodies. Congressalso authorized EPA to set the criteria if States failed to do so by February 1990 or todevelop replacement standards if the agency believes a State’s standards do not meetminimum requirements (Copeland 1993; Fogarty 1991). In practice, EPA has permittedthe States some discretion in developing their criteria. Under EPA’s 1983 revisions towater quality regulations, States retain the right to modify EPA criteria to reflect site-specific conditions or adopt numerical values based on “other scientifically defensiblemethods” (Executive Enterprises 1984, citing 40 CFR 13.11(b)(1)).
In 1990, for example, the State of Maryland proposed a water quality standard fordioxin 10-fold higher than EPA’s numeric criteria based on an allowable one in onehundred thousand (10-5) cancer risk. By EPA’s reckoning, Maryland’s proposed standardsuggested a cancer risk (10-4) of potential concern. Acknowledging that there are avariety of equally defensible scientific assumptions that can be made, however, the agencyapproved Maryland’s standard. In the State’s proposal, many of the scientific assumptionswere the same as those of EPA; where they differed (e.g., the estimated carcinogenicpotency of dioxin), Maryland used alternative assumptions employed by the Food andDrug Administration (Moore et al. 1993; Thompson and Graham 1997). Thus, the CleanWater Act is unusual among federal environmental statutes in the extent to which EPAand the States share authority to set risk-based public health standards.
Section 304 of the 1987 Clean Water Act Amendments directed States to developlists of their impaired waters by 1989. Impaired waters are those bodies that do not meetor are not expected to meet ambient water quality standards, even after implementation oftechnology-based controls implemented by point sources. The States were also requiredto identify point sources causing the water quality impairments and develop individualcontrol strategies to control those sources further (Copeland 1993). Under the CWA,developing these controls is to be done by setting the total maximum daily load (TMDL),the maximum quantity of a pollutant a water body can receive daily without violatingambient water quality standards under local conditions. The TMDL is then to be allocatedamong the various sources contributing to the problem. Finally, the National PollutionDischarge Elimination System (NPDES) permits for regulated point sources are to berevised, as warranted.9 If the States failed to identify a list of impaired waters and develop 9 Although non-point sources such as runoff from most farms and roadways and atmospheric depositionmay contribute to exceedances of ambient water quality criteria, enforceable limits can only be placed on
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 6
TMDLs, the 1987 CWA Amendments required EPA to develop a priority list for the Stateand make its own TMDL determination. In response, EPA mandated that the list ofimpaired waters include those receiving discharges from pulp and paper mills and calledfor specific limits on dioxin discharges by 1992 (Thompson and Graham 1997).
In 1990, at the request of Oregon, Washington, and Idaho, EPA established itsfirst TMDL for eight pulp and paper mills discharging into the Columbia River Basin(which includes the Snake and Willamette Rivers). Each of the States had adopted thesame ambient water quality standard for dioxin (0.013 parts per quadrillion (ppq)).10
Based on considerations of regional hydrology, other sources of dioxin, etc., EPA set aTMDL for dioxin (of 5.97 milligrams per day) and allocated 35% of the load to U.S. pulpand paper mills operating in the river basin. Environmental groups sued EPA for notsetting a more stringent TMDL, and the pulp and paper mills sued the agency for settingthe TMDL before finalizing new effluent guidelines for the entire industry (Thompson andGraham 1997). In 1995, the U.S. Ninth Circuit Court of Appeals upheld EPA’s TMDLfor dioxin in the Columbia River basin (Environment Reporter, 6/30/95, p. 493).
In general, the impaired waters listing process/TMDL program has labored underthe Clean Water Act’s system of shared EPA-State responsibility. With the CWArequiring EPA to serve as a backstop, state environmental agencies may have littleincentive to allocate limited resources to the program and take the heat for controversialdecisions. The TMDL program has come under increased fire from environmental groups,tribes, industry, and local communities. A series of recent court decisions citing EPA’sfailure to complete the tasks after States failed to do so within the statutory time limitscould force the agency to make an incredible number of geographically-specificdeterminations under demanding time, data, and resource constraints. (In the State ofIdaho alone, for example, a federal district court has required EPA to set TMDLs for over900 water segments in a five-year time period (Inside EPA, 10/4/96, p. 4).) For EPA andstate environmental agencies, the analytically and politically daunting task of setting andallocating innumerable TMDLs makes it all the more appealing to formulate national,technology-based effluent guidelines so as to limit the number of water bodies expected toexceed ambient water quality standards. Environmentalists seek to avoid the cost anddelay involved in case-by-case regulation and are wary that States may be reluctant toimpose additional controls on firms within their borders. Individual firms or plants alsohave an interest in assuring that geographically-specific pollution controls do not put themat a competitive disadvantage. Thus, the CWA provisions requiring EPA and the Statesto consider geographically-specific conditions may influence the use of scientificinformation in national rulemaking.11
point sources. Non-point and mobile sources may contribute to background levels of dioxins and otherorganochlorines that end up in surface waters, sediments, and aquatic organisms.10 1 ppq is 1 x 10-15 .11 In 1996, EPA began drafting a strategy to administratively reform the TMDL program and convened aFederal Advisory Committee Act group to develop recommendations. (See Inside EPA, 11/22/96, pp. 4-6for a summary of the agency’s draft strategy).
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 7
Mirroring the shared regulatory authority between EPA and the States, the CWAis implemented in an environment in which Congress and the Executive Branchcontinuously wrestle for control over regulatory policy. Spurred by a series of ExecutiveOrders dating back to the Nixon administration requiring some form of economic analysisfor proposed regulations (but in particular, the 1981 Reagan administration ExecutiveOrder 12291 requiring Office of Management and Budget (OMB) review of newregulations), EPA has deemed it “appropriate” to consider cost-effectiveness comparisonswhen proposing BAT for toxic water pollutants (see discussion of Sec. 304(b)(2) above).According to sources in the EPA Office of Water, however, the program generally regardscost-effectiveness analysis as an imprecise tool that only permits a rough screening ofregulatory options, and the agency has not explicitly made any BAT decisions on the basisof cost-effectiveness.
Thus, while the CWA prods EPA to do what is “doable” to reduce toxic waterpollution, OMB pulls the agency toward what it thinks is “reasonable.” As discussed ingreater detail below, the projected benefits of the nominally “technology-based”regulations to limit dioxin and other organochlorines from pulp and paper mills areestimated using the tools of environmental science and risk assessment. Disagreementsabout the agency’s regulatory proposals are often conducted in the language of scienceand technology and are, in part, over how to properly assess its environmental benefits.The subtext, however, is whether those benefits are reasonably associated with compliancecosts.
In addition to its prominent role in the proposed pulp and paper cluster rule, dioxinhas a long and highly publicized history. As Finkel (1988) noted, our nationalpreoccupation with dioxin stems largely from the notoriety of TCDD as the most potentanimal carcinogen ever tested, and its ubiquity as a contaminant of pesticides, incineratorsmoke and ash, and bleached paper consumer products such as diapers and coffee filters.More recently, the dioxin story has segued into the broader debate over “endocrinedisruptors,” a class of hormone-like chemicals suspected of having a variety ofreproductive and other non-cancer effects. Endocrine disruptors are the subject of themuch-discussed popular science book entitled, Our Stolen Future, which argues thatbackground levels of chlorinated organics and other industrial chemicals may play a role indevelopment of breast cancer, falling sperm counts and other male reproductive disorders,as well as developmental effects in wildlife and humans (Colborn et al. 1996).12
Forty years ago, a European researcher identified the impurity TCDD as causingthe skin disease chloracne in chemical workers involved in the production of the herbicide2,4,5,-T (Moore et al. 1993). But dioxin first came to public light in the early 1970s as aresult of concerns about the exposure of Vietnam Veterans and South Vietnamese children
12 See Hirshfield et al. (1996) for a thoughtful review of Our Stolen Future and comparison to RachelCarson’s 1962 Silent Spring, which publicized the environmental effects of pesticides and is associatedwith the birth of environmentalism as a mass movement.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 8
to the defoliant Agent Orange (which included 2,4,5,-T).13 EPA promulgated a partial banon the herbicide in 1971. The animal studies that resulted in dioxin (TCDD) being labeledas the most potent carcinogen were conducted in 1978. One year later, EPA issued acontroversial reanalysis of an epidemiological study conducted in Alsea, OR (Alsea II)which associated miscarriages with herbicide spraying, leading to accusations that EPAhad “cooked” the data to inflate the risks (Whelan 1985), and the agency suspendedessentially all remaining uses of 2,4,5-T.14 The problem of dioxin emissions frommunicipal waste incinerators was identified in 1979 and gained public notoriety as themain plank of Barry Commoner’s 1980 presidential campaign platform. (Later, debatesover the location and siting of incinerators gave impetus to the environmental justicemovement.)
In 1974, the federal Centers for Disease Control (CDC) identified dioxin as a toxicsubstance in Missouri waste oil. In 1982, EPA detected high dioxin levels from TCDD-contaminated oil sprayed on streets in Times Beach, MO, and dioxin was implicated inillnesses in horses and possibly children. Flooding in December raised concerns aboutcontamination spreading to other sites (though it did not) (OTA 1991). In 1983, the CDCand the Missouri Division of Health recommended that the town be evacuated, and EPAand the Missouri Department of Natural Resources paid $36 million to buy all 801 homesin Times Beach and relocate its residents because of the unavailability of demonstratedtreatment technologies and the uncertainty about when the cleanup would be completed.A $200 million cleanup of the town's 400 deserted acres was later initiated. In the summerof 1990, Vernon Houk, head of the Center for Disease Control's Center for EnvironmentalHealth and Injury Control, told a congressional committee that new evidence suggestedthe risk of dioxin historically was vastly overstated.15
EPA’s first health assessment of dioxin was conducted in 1981 and was revised in1985. Animal studies by Dow Chemical Co. researchers (Kociba et al. 1978) and theNational Toxicology Program (NTP 1982) were important sources of scientificinformation for the agency’s assessment. The 1985 assessment is the current official basisof dioxin cancer risk estimates used by EPA for all regulatory decisionmaking, includingthe 1993 proposed pulp and paper cluster rule. However, EPA has been in the process ofreassessing the risks of dioxin for several years. During the 1980s, some researchers
13 A 1969 National Cancer Institute study found a link between TCDD and birth defects. According toSmith (1992), other studies by U.S. scientists critical of the Vietnam War also reported teratogenic effectsof TCDD. Restrictions on domestic uses of 2,4,5,-T were first announced in 1970 by the Secretary ofAgriculture. In promulgating a partial ban on the herbicide in 1971, EPA Administrator WilliamRuckelshaus rejected the advice of an ad hoc scientific panel chaired by Emil Mrak, Chancellor ofUniversity of California, Davis, and accepted the counsel of a group Food and Drug Administration(FDA) scientists who had conducted earlier animal tests on 2,4,5,-T. Critics of the Mrak panel hadreceived leaked copies of the report prior to its release. Both advisory groups were informally convenedprior to the advent of the 1972 Federal Advisory Committee Act (FACA), and the episode crystallizedsupport for FACA (Smith 1992, p. 24-25).14 See Jasanoff 1990, pp. 24-26 for a more balanced discussion of the 2,4,5-T controversy.15 Two sources interviewed for the overall study of science in environmental regulation volunteered theTimes Beach buy-out as an EPA decision in which science played little or no role.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 9
postulated that dioxin might “promote” rather than “initiate” cancer and that, as a result,EPA may have overestimated the cancer risks from dioxin. With the backing of AssistantAdministrator for Pesticides and Toxics John Moore, an ad hoc EPA committee in 1986recommended moderating the dioxin cancer risk estimate. At about the same time, OMBhighlighted the large scientific uncertainty of dioxin cancer risk estimates in its annualreport on federal regulatory programs (Roberts 1991; Moore et al. 1993). The followingyear, EPA issued a draft reassessment suggesting that the risk of cancer from dioxin was17 times less than the agency had assumed. According to Finkel (1988), however, theagency developed its revised estimate not on the basis of any new data, but by essentiallysplitting the difference between two “fundamentally irreconcilable theories about thecarcinogenicity of dioxin.” Regardless of whether the decision was “right for the wrongreasons,” as some felt, the agency’s approach could not withstand scrutiny. In reviewingthe agency’s draft, the EPA Science Advisory Board (SAB) criticized EPA’s currentcancer risk assessment methodology but found no new data to support changing the dioxincancer risk estimate (Moore et al. 1993).
In 1990, Robert Scheuplein of the Food and Drug Administration (FDA),toxicologist Michael Gallo of the Robert Wood Johnson Medical School in New Jersey,and Dutch scientists organized the “Banbury Conference,” (held at New York’s ColdSpring Harbor Lab) which formally marked a new scientific consensus about a series ofbiological steps occurring at the molecular level that precede most if not all of theobserved effects of dioxin and other similar chemicals (Roberts 1991).16 Some scientistsinterpreted this to mean that the very low levels of dioxin in the environment would resultin negligible cancer risks. In 1991, an epidemiological study conducted by NationalInstitute of Occupational Safety and Health (NIOSH) researchers (Fingerhut 1991)reported a statistically significant increased cancer risk in U.S. chemical workers exposedto high levels of dioxin but detected no increase in workers exposed to low levels. As aresult of the Banbury Conference and the NIOSH study, external pressures mounted forEPA to move beyond research and initiate a formal reassessment of dioxin. According topress reports, the paper industry was a leading voice in persuading the agency to revisitdioxin (Rachel’s Environment & Health Weekly, 8/31/95, p. 1). In April 1991, EPAAdministrator William Reilly announced that the agency would comprehensively reassessthe cancer and non-cancer risks of exposure to TCDD and related compounds.
While EPA slightly moderated the cancer risk estimate for dioxin and similarcompounds in its draft reassessment released in 1994, it also concluded that there waspotential for a variety of adverse non-cancer effects in the range of current backgroundexposures to dioxin and similar compounds (EPA 1994b). In reviewing the draftreassessment, a majority of SAB members concluded that agency tends to overstate thepossibility for danger at near-ambient levels, but several SAB members regard theagency’s characterization of the risks as appropriately conservative within the context of
16 Consensus broke down, however, on just what such a biologically-based model would predict in termsof dioxin’s cancer risks (Roberts 1991). See Powell (1996) for a discussion of Gallo’s role in promotingEPA’s use of biologically-based risk assessment models. According to a former senior EPA official,industry, notably the Chlorine Institute, played a role in initiating the Banbury Conference.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 10
public health protection (EPA/SAB 1995). An environmentalist now says, “Reilly’sdecision to conduct the dioxin reassessment did not turn out the way he dreamed it would.Industry and he thought they would have a slam dunk on dioxin’s carcinogenicity.”
Although it appeared likely to many in 1988 when EPA began to formulate thenew pulp and paper effluent limits that our “national preoccupation” with dioxin wouldwane, the agency’s subsequent dioxin reassessment has highlighted the non-cancer effectsof dioxins and helped launch the issue of endocrine disruptors onto the environmentalregulatory agenda. Some environmentalist groups (notably Greenpeace) have respondedby calling for a ban on chlorine. This proposal was afforded a measure of mainstreamlegitimacy in February 1994 when the International Joint Commission (IJC), the Canadian-American bilateral organization established to monitor the Great Lakes Water QualityAgreement, recommended phasing out the use of chlorine and chlorine-containingcompounds as industrial feedstocks. It is in this context that EPA will try to finalize theeffluent regulations for the pulp and paper industry. Table A-1 provides a summarybackground of dioxin science and policy. Table A-2 summarizes the development of thepulp and paper cluster rule.
Table A-1. Summary Background of Dioxin Science and Policy
1949 USDA registers 2,4,5-T as a pesticide.1957 TCDD identified as causing chloracne.1966 USDA and FDA establish residue tolerances for 2,4,5-T in food.1969 Initial laboratory studies link 2,4,5-T and TCDD with birth defects.1970 U.S. halts use of Agent Orange in Vietnam.1971 EPA restricts domestic use of 2,4,5-T.1972 Controversy over EPA’s 2,4,5-T decisionmaking process crystallizes congressional support for Federal
Advisory Commission Act.1974 CDC identifies dioxin as toxic substance in Missouri waste oil.1976 Industrial accident releases large quantities of dioxin in Sveso, Italy.1977 Clean Air Act Amendments list dioxins and furans as hazardous air pollutants.1978 First EPA study regarding linkage between miscarriages and herbicides in Alsea, OR.
Dow Chemical Co. researchers report that TCDD is a carcinogen in laboratory studies.1979 Alsea II reevaluates miscarriage-herbicide data. EPA accused of inflating risks.
EPA suspends use of 2,4,5,-T. Vietnam veterans start class action suit.Dioxin and furans identified in emissions from municipal waste combustion plants.
1980 Barry Commoner’s presidential campaign elevates concerns about dioxin releases from wasteincinerators.
1981 Sveso 5-year report finds no dioxin effects other than chloracne.EPA’s Cancer Assessment Group estimates that dioxin is one of the most potent carcinogens known.
1982 National Toxicology Program reports results of dioxin animal cancer study.1983 Times Beach, MO buyout.
EPA issues congressionally-mandated national strategy to investigate, identify, and remediate dioxincontaminated areas.
1984 EPA cancels 2,4,5-T registration.Hazardous Solid Waste Act requires EPA to evaluate risks posed by dioxin emissions from municipalwaste combustion facilities.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 11
Table A-1. Summary Background of Dioxin Science and Policy (cont’d)
1985 EPA revises its dioxin health assessment, lowering the cancer risk estimate by more than a factor of 2,but retains the agency’s default linear cancer model.
1986 Ad hoc expert committee advises EPA that linear cancer model is inappropriate for dioxin.1987 EPA scientific group recommends moderating cancer risk estimate.
EPA develops Toxic Equivalency Factors (TEFs) for dioxin and dioxin-like chemicals.1989 EPA SAB finds no new data to support change in cancer risk estimate; critical of current cancer model;
accepts TEFs as an interim approach.1990 Banbury Conference supports receptor-mediated event for dioxin activity.
EPA promulgates New Source Performance Standards for municipal waste combustion facilitiesrequiring best management practices to limit total dioxins and furans to 30 ng/m3.
1991 NIOSH epidemiological study suggests that dioxin is a human carcinogen, but perhaps only at highlevels of exposure.EPA initiates dioxin reassessment.
1994 EPA draft dioxin reassessment reports potential for adverse non-cancer health effects within the rangeof current background levels.Chlorine ban proposed by Henry Waxman (D-CA), Barry Commoner, and others (EnvironmentReporter, 9/30/94, p. 1133).
Table A-2. Development of the Pulp and Paper Cluster Rule.
1983 EPA initiates national dioxin survey, detects elevated dioxins downstream from pulp andpaper mills.
1984 EDF and NWF file TSCA petition requesting EPA to regulate dioxins and furans from allknown sources. EPA denies petition.EPA issues Ambient Water Quality Criteria report for dioxin.
1985 EDF and NWF file lawsuit.1986 June. EPA, NCASI, and American Paper Institute (API) agree to undertake the “5 Mills
Study,” detect TCDD and TCDF in effluents, pulp and sludges of pulp and paper mills.December. Information on the agreement between EPA and the pulp and paper industryreported. Greenpeace initiates Freedom of Information Act (FOIA) request seeking allavailable information on the pulp mill dioxin problem.
1987 Clean Water Act Amendments establish deadlines for EPA and States to address toxicpollutants.January. Letter from EPA to API leaked to environmentalists indicates EPA officials hadagreed to notify the industry “immediately” of receipt of any requests under FOIA and that,barring such requests or results indicating a potential threat to human health, the agency didnot intend to release any results until publication of the final report on the study.August. Greenpeace USA releases report alleging an EPA cover-up.September. New York Times front-page story reports traces of dioxin detected in householdpaper products. Report based on the “5 Mills Study” and analyses of dioxin in paperproducts.
1988 EDF, NWF and EPA sign consent decree requiring agency to perform a comprehensive riskassessment of dioxins and furans considering sludges, water effluent, and products madefrom pulp produced at 104 bleaching pulp mills and (as amended in 1992) to proposeregulations addressing discharges of dioxins and furans into surface waters from the mills byOctober 31, 1993.EPA issues “interim strategy” to address dioxin emissions from pulp mills, which includedrequiring pulp mills to monitor for dioxins and adopt short-term control measures (Hanmer1988; EPA-V 1988).EPA and industry begin the “104 Mill study.”
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 12
Table A-2. Development of the Pulp and Paper Cluster Rule (cont’d).
1988 Swedish studies generate adsorbable organic halides (AOX) indicator used in EPA’s 1993proposed effluent limits.
1989 EPA initiates inter-agency, inter-office assessment of pulp and paper sludges, effluents, andconsumer products.OTA report discusses Swedish pulp mills’ compliance with more stringent regulatorystandards for organochlorine emissions (OTA 1989).March and June. First results of “104 Mill study,” released.
1990 EPA issues Assessment of Risks from Exposure of Humans, Terrestrial and Avian Wildlife,and Aquatic Life to Dioxins and Furans from Disposal and Use of Sludge from BleachedKraft and Sulfite Pulp and Paper Mills. Based on 104 Mill study, assessment estimates thatpreventing adverse wildlife effects would require TCDD soil concentrations 4-400 timeslower than levels needed to prevent unacceptable human health risks.
1991 May. Under court consent decree, EPA proposes pulp and paper mill sludge rule underTSCA Sec. 6. Proposal would set a 10 ppt maximum allowable dioxins/furans concentrationfor land application (resulting in an estimated human health risk less than 10-4) and includesprovisions for mills to submit annual reports and maintain records on land, application, andlaboratory analysis.July. OMB objects to proposal’s information collection request (Environment Reporter,8/16/91, p. 1058).
1992 EPA announces it would seek a voluntary agreement with industry on the pulp and papermill sludge rule (Environment Reporter, 12/24/93, pp. 1545-1546).
1993 September. NRDC and 55 other environmental groups petition under CWA Sec. 307 (a) forEPA to ban dioxin discharges by the pulp and paper industry by prohibiting the use ofchlorine rather than manage dioxin through BAT standards under pulp and paper clusterrule (Environment Reporter, 9/17/93, pp. 889-890).December. EPA proposes pulp and paper cluster rule based on BAT standards.
1994 February. At hearing on proposed cluster rule, industry representatives claim that EPA’senvironmental benefits analysis does not employ sound science and overstates benefits.Future EPA Assistant Administrator for Research and Development Robert Huggett reportsthat substitution of chlorine dioxide for elemental chlorine reduces chemicals thataccumulate in fatty tissues to the limits of detectability (Environment Reporter, 2/18/94, pp.1783-1784).April. EPA and pulp and paper industry announce voluntary agreement regarding landdisposal of dioxin-tainted sludge formalizing best management practices. No restrictions onuse of sludges if concentration of dioxin and furan is less than 10 ppt. For pasture lands, theconcentration limit is 1 ppt (i.e., background levels). At 50 ppt, sludge cannot be landapplied.
1996 On the basis of new data regarding the environmental performance of pulp and paper millsthat have completely substituted chlorine dioxide for elemental chlorine, EPA announcesthat it is considering two BAT options for the major pulp and paper subcategory (bleachedpaper papergrade kraft and soda).
2. Scientific Issues
Dioxin and Related Compounds
The major scientific controversy over dioxin and its chemical cousins is notwhether high levels of exposure can cause cancer in humans but rather the risks posed by
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 13
background levels and incremental releases of all dioxin-like compounds. Althoughdioxins and other organochlorines have been associated with a variety of non-cancereffects, the conventional focus of scientific investigation has been on cancer. According toLucier et al. (1993), several long-term bioassays have been conducted on TCDD inseveral species. All studies have produced positive results. It is clear the TCDD is amultisite carcinogen in both sexes of rats and mice. It also is a carcinogen in the hamster,which is considered the most resistant species to the acute toxic effects of TCDD. TCDDis also found to increase cancer incidence in animals at doses well below the MaximumTolerated Dose. While TCDD appears to be a chemical that strongly promotes cancerdevelopment once initiated, it seems to have weak or no potential to initiate cancer itself.The general consensus is that TCDD is an example of a carcinogen whose action ismediated by a specific receptor within cells, suggesting that there may be a threshold dosebelow which dioxin is not carcinogenic. It may be possible, however, that non-cancerhealth effects result at levels below the threshold dose for cancer.
A considerable body of studies of people exposed to dioxin provides suggestiveevidence of its human carcinogenicity, but according to an EPA official, theepidemiological evidence is inconclusive due to a number of factors. First, scientistscannot be certain about how much dioxin and other chemicals the subjects were exposedto. Second, in most studies, the numbers of people exposed through accidents or in theworkplace have been too small to allow scientists to detect substantial changes in cancerrates. Third, those individual who were exposed to dioxin (mostly healthy adult males)may not have been the most sensitive group. Finally, not enough time may have elapsedbetween exposures and study completion for most cancers to develop (many cancers onlydevelop 15-30 years after exposure). The first dioxin epidemiological study sufficientlylarge enough to detect a substantial increase in cancer doses, according to this EPAofficial, was Fingerhut et al. (1991). This NIOSH study, which took nearly 13 years tocomplete and examined 5172 male U.S. chemical workers exposed to dioxin on the jobfrom 1942-84 presented what many consider the strongest evidence that dioxin is a humancarcinogen--but perhaps only at very high doses (Roberts 1991). The EPA ScienceAdvisory Board has agreed that although human data are limited, dioxin is a probablehuman carcinogen under some exposure conditions (EPA/SAB 1995). In February 1997,an International Agency for Research on Cancer (IARC) Working Group also concludedthat TCDD should be considered carcinogenic to humans(http://www.iarc.fr/preleases/115e.htm).
Extrapolating from rodent studies using a linear model of cancer risk, EPA’sCancer Assessment Group derived an extraordinarily high cancer potency factor (4.25 x105 (mg/kg/day)-1) for dioxin in 1981. An important basis of this estimate was a reanalysisof the pathological evidence from the Dow Chemical researchers’ rat study (Kociba et al.1978) performed by Robert Squire of Johns Hopkins University Medical School.17
17 Pathology includes laboratory analysis of animal tissue slides to characterize and enumerateabnormalities such as tumors. It is traditionally descriptive and can be fairly imprecise, but standardizedprotocols and quantitative and chemical techniques have been developed to promote consistency andprecision.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 14
Squires’ reinterpretation of the tissue samples resulted in a cancer potency factorapproximately two times higher than the one derived using the original diagnoses. In1985, EPA revised its dioxin cancer potency estimate downward by more than a factor oftwo (to 1.56 x 10-5 (mg/kg/day)-1) by adjusting for the early mortality of study animalsobserved in Kociba et al. (1978) and by essentially splitting the difference (taking thegeometric mean) between the original pathology assessment and Squires’ reanalysis(Thompson and Graham 1997). The agency had moderated its dioxin hazard assessmentsomewhat, but it still estimated that a one in a million (10-6) cancer risk was associatedwith exposure to the infinitesimally small quantity of 0.006 pg/kg/day (picograms (10-12 g)per kilogram body weight per day).
Although EPA indicated in 1985 that there was inconclusive evidence that dioxinwas a mutagen (able to initiate carcinogenisis), the agency determined that the availabledata on dioxin’s biological activity (carcinogenic mechanism and pharmacokinetics) wereinsufficient to support deviation from the default linear dose-response model for cancer.Canada and European countries, however, rejected the linear cancer model asinappropriate for dioxin because it is not considered genotoxic (i.e., dioxin does notdirectly initiate cancer by causing mutation or DNA damage), and set their limits at 1-10pg/kg/day. There were also differences in dioxin cancer potency estimates within the U.S.government between EPA, CDC, and FDA. FDA’s cancer potency estimate is almost anorder of magnitude smaller than EPA’s 1985 estimate, and CDC’s is intermediate betweenthe two. The inconsistent estimates resulted from the agencies applying the same linearcancer model but making a variety of different scientific assumptions and datatreatments.18
In the 1970s, Alan Poland of the University of Wisconsin initiated the first studieson dioxin’s biological mechanisms (Thompson and Graham 1997). At the 1990 BanburyConference, scientists agreed that the biological activity of dioxin and dioxin-likecompounds was mediated by first binding to a specific molecular receptor in cells, the arylhydrocarbon (Ah) receptor (an intracellular protein).19 Theoretically, dioxin moleculesmay have to occupy many Ah receptors sites before any biological response is seen, andeven once activity begins, the cell’s internal regulation system has some capacity to adaptto changing hormonal levels and maintain the mix within the range of tolerance. In theview of some scientists, this theoretical argument suggests a threshold below which dioxincannot cause cancer and implies that EPA’s linear cancer model is invalid for dioxin. “If
18 The agencies’ estimation procedures differed in how to extrapolate from rat to man (body weight orsurface area); which pathology results were used (Kociba and colleagues’, Squire’s, or both); whetherearly mortality was taken into account; the assumed average human body weight (80 kg or 70 kg); andhow the dose was measured (concentration in the tissue or administered dose) (Thompson and Graham1997). Using surface area to scale the administered dose between animals and humans leads to a higherpotency estimate than does using body weight as a scaling factor. Currently, EPA uses a scaling factor ofbody weight raised to the 2/3 power. According to an academic, there is a proposal for all federalagencies to adopt a scaling factor of body weight raised to the 3/4 power, but FDA continues to scale onthe basis of body weight.19 In 1995, the EPA Science Advisory Board reported that it was also possible that dioxin may producetoxic responses that are not mediated through the Ah receptor (Thompson and Graham 1997).
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 15
we can’t do it [depart from the linear default model] for dioxin, for which we have somuch information, then we probably can’t do it for anything,” said Banbury Conferenceorganizer Robert Gallo (quoted in Roberts 1991).20
However, there may be considerable variability among individuals in the thresholdlevel at which carcinogenesis begins. In addition, a continuum of biological activity occursbeginning at relatively low levels of Ah receptor occupancy. There is, however,considerable controversy regarding the health significance of the activities initiated atlower levels of occupancy. (In practical terms, this means that setting dioxin limits lowenough to prevent cancer may be insufficient to prevent other biological effects, but the“so what?” question has yet to be resolved by scientific consensus.) In response to the1991 decision to conduct a dioxin reassessment, scientists at EPA’s Office of Researchand Development and the National Institute of Environmental Health and Safety (NIEHS)began research to characterize a threshold for dioxin in humans. The results, reported in1992-93, suggested that enzyme induction occurs at existing background levels of dioxin-like compounds (Thompson and Graham 1997). Instead of cancer being initiated at thelowest dose levels, it is now hypothesized that reproductive, developmental, and immune-system impairments may be the most sensitive health effects of dioxin. For these non-cancer effects, says an environmentalist, the old toxicological adage that “the dose makes
21 Instead, the timing--not the quantity--of exposure may bethe critical factor. This source is concerned, for example, that exposure to a trace quantityof dioxin that might be irrelevant in terms of cancer risk could result in a substantialdevelopmental risk if maternal exposure occurs at a critical period of fetal development.
Further, the biological system responds to the cumulative exposure of dioxin andsimilar chemicals that bind to the Ah receptor rather than to the exposure to any singledioxin-like compound. As a result, much disagreement now centers on just how closeexisting background levels of all dioxin-like compounds occurring in the environment andstored in human tissues are to the levels required to cause adverse health effects.22 AsThompson and Graham (1997) suggest, the significance of this dispute is that the conceptof a threshold level of Ah receptor occupancy may be irrelevant to decisions aboutadditional releases of dioxin-like compounds if typical body burdens already exceed thethreshold.
Of the group of 75 chlorinated dioxins, only TCDD has been subjected to long-term animal carcinogen experiments. To account for the cumulative exposure tocompounds that, like dioxin, would bind to the Ah receptor, in 1987, the EPA RiskAssessment Forum developed Toxic Equivalency Factors (TEFs). These TEFs derivefrom a relative ranking scheme based on assigning a TEF of 1.0 to TCDD, since it showsthe greatest affinity for binding to the Ah receptor. Other dioxin-like compounds are
20 Similar statements have been made regarding departure from the linear model for ingested arsenic. SeePowell (1996).21 This means that too much of anything--even something essential to life in normal doses--can beharmful.22 Background levels would include accumulations of both natural and anthropogenic sources.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 16
assigned a fractional weight proportional to their binding affinity relative to that of TCDD.The TEFs are intended to be additive weighting factors. The TEF for TCDF, for example,is 0.1--its affinity for binding to the Ah receptor is 1/10th that of TCDD (EPA 1989).(Thus 5 g of TCDD plus 5 g of TCDF yields the estimated equivalent of 5.5 g of TCDD.)There is not a perfect correlation, however, between Ah receptor binding affinity and thepotency for various toxic effects. Consequently, there is considerable uncertainty abouthow accurately TEF equivalent weights reflect cumulative effective exposures.23
As indicated earlier, dioxins are produced in very small quantities. EPA (1994a)estimates annual emissions from known sources for the entire U.S. at 3,300 - 26,000grams, with the total possibly being as high as 50,000 g/yr.24 However, dioxins areextremely insoluble in water, environmentally and biologically stable, persist in theenvironment for long periods, and tend to accumulate in animal tissues. Thus, thepredominant route of human exposure is probably through the food chain rather thaninhalation or drinking water. Relative to other foods, measurements of background levelsof dioxin are particularly high in fish. Currently, bleaching pulp and paper mills are theonly significant known source of dioxins released into surface waters (EPA 1994a).According to an EPA official, the agency estimated a very wide range of risks resultingfrom dioxin and furan released from pulp and paper mills, much of which was explained bythe size of the receiving water body into which plant effluent was being discharged.
Formation of Dioxin and other Organochlorines from Bleaching Pulp
Lignin is a natural polymer that binds and supports cellulose fibers of woodyplants, but it discolors and weakens paper products. Chemical pulping dissolves a largefraction of lignin using nonoxidizing chemicals (e.g., alkalis or sulfites) while preserving alarge fraction of the desired cellulose fibers. Various forms of chlorine and otherbleaching agents are used to further remove lignin from pulp to produce durable whitepaper products (like this page). For many decades, elemental chlorine (Cl2) has been thebleaching agent of choice for much of the U.S. pulp and paper industry due to its relativelylow cost. Chlorine dioxide (ClO2) is more selective for lignin and thus can achieve thesame level of pulp bleaching with a substantially lower input or “charge” of chlorine, but itcosts more than elemental chlorine. Using a process called oxygen delignification (OD),oxygen may also be used as an initial bleaching agent to reduce the chlorine charge
23According to a former Science Advisory Board member, the Environmental Defense Fund encouragedEPA to develop the TEF scheme. When the Board reviewed the scheme in the late 1980s, says thissource, “The SAB said, ‘We’ll accept that as an interim procedure, but more research is needed tosubstitute for TEFs.’ Now the TEFs are getting locked in, and the research wasn’t done. People get usedto using the old numbers, and they take on a life of their own. There’s a ‘check the box’ mentality, aresistance to revisiting old decisions. Risk assessment needs to be an iterative process.”24 These figures are for all dioxin-like compounds weighted by toxic equivalency factors, but they aredominated by TCDD (about 90% of the total).
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 17
required to achieve a given level of pulp brightness. However, OD is a capital-intensivetechnology.25
When dioxin was first detected in streams below pulp and paper mills, the firstculprits identified were oily defoamers and woodchips treated with polychlorophenols.Addressing these sources, however, did not eliminate dioxin formation from bleachingpulp and paper mills. This suggests that some dioxin and furan precursors might occur intrees naturally (Berry et al. 1991). It now appears that the only way to entirely preventformation of dioxins, furans, and other organochlorines by the pulp and paper industry isto eliminate the use of chlorine as a bleaching agent. By substituting the more lignin-selective ClO2 for elemental chlorine, however, the formation of organochlorines--andparticularly the persistent, bioaccumulable polychlorinated organics or greatest concern--can be dramatically reduced.
According to Berry et al. (1991), of the chlorine used in pulp bleaching, about90% ends up as common salt (e.g., calcium chloride) and about 10% binds to organicmaterial removed from the pulp. About 80% of this organically bound chlorine occurs inhigh-molecular weight material that does not permeate cell walls and is relatively watersoluble.26 Most of the organically bound chlorine which occurs in low-molecular weightcompounds that can permeate cell walls is relatively water soluble and is readilyhydrolyzed or metabolized. A small fraction (about 1%) of the total organically boundchlorine is relatively fat soluble and potentially bioaccumulable and toxic. A component ofparticular concern in this fraction is the polychlorinated organic material, which includesdioxin, furan, and polychlorinated phenolic compounds. The polychlorinated phenoliccompounds, however, are considered much less toxic than dioxin. For example, EPAestimates the cancer potency of 2,4,6-Trichlorophenol to be seven orders of magnitudelower than that of TCDD (EPA 1993a, Table 3-1).
Because chlorine atoms are added to organic precursors in a largely sequentialprocess (with the di-chlorinated organics most likely to be formed before tri-chlorinatedorganic, tri-chlorinated organics most likely to be formed before tetra-chlorinatedorganics, etc.), Berry et al. (1991) concluded that a threshold level of chlorine chargewould be required for any TCDD and TCDF formation to occur. They further suggestedthat 100 percent substitution of ClO2 for Cl2 (called “complete substitution”) couldprevent such formation. However, more recent data from mills employing completesubstitution show detectable levels of TCDD and TCDF in bleach plant effluents (ERG1996). Given the huge number of randomly interacting molecules present in commercial- 25 Because oxygen is relatively unselective for lignin, OD results in more dissolved organic material.Extended cooking has a similar effect. Consequently, pulp and paper mills using these delignificationtechnologies require more recovery boiler capacity than those mills that do not.26 Berry et al. (1991) surmise that it is highly improbable that the high-molecular weight chlorinatedlignin material would be broken down and transformed in the environment into problematic,polychlorinated compounds because the potentially troublesome aromatic (6-carbon ring) structure of theresidual lignin would largely be destroyed by oxidation in the bleach plant. Berry et al. (1991) add,however, that further investigation of the environmental fate of this fraction of the organochlorines isneeded to confirm that neither it nor its decomposition products are harmful.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 18
scale pulp bleaching, one would expect some trace (perhaps undetectable) amounts of tri-chlorinated organics (such as trichlorophenol) and tetra-chlorinated organics (such asTCDD and TCDF) to be formed at even the lowest chlorine charges, particularly if thepulp and chlorine are not uniformly mixed.27 Thus, complete substitution of ClO2 forelemental chlorine would not entirely eliminate dioxin and furan formation. Completesubstitution does appear, however, to reduce dioxin and furan formation to the flat portionof the curve, well beyond the point of diminishing returns. (See the data presented inBerry et al. 1991.)
Berry et al. (1991) also observed that the formation of dioxin and furan is littleaffected by the lignin content of unbleached pulp. This conclusion has been reinforced bythe more recent environmental performance data. TCDD and TCDF were not detected inany industry-supplied sample results from bleached papergrade kraft mills employingcomplete substitution (of Cl2 with ClO2). But TCDD and TCDF were detected in EPA-collected samples at several mills using both complete substitution and oxygendelignification (ERG 1996). Therefore, while complete substitution may not entirelypreclude dioxin formation, initiating the bleaching process with OD (and thereby furtherreducing the required chlorine charge) does not appear to prevent it either. In contrast,the lower lignin content of pulp prior to bleaching plays a decisive role in the reducedformation of the less toxic but more abundantly formed chlorinated phenolics (Berry et al.1991).
Releases and Detection of Dioxin and other Organochlorines from Bleaching Pulp
For contaminants like dioxin that are toxic at trace concentrations, damages mayoccur at environmental levels resulting from the cumulative releases of multiple sourceswhich, when considered individually, may discharge undetectably low concentrations ofthe pollutant. EPA’s 1993 proposed BAT for pulp and paper effluents was expected toyield non-detectable concentrations of TCDD for two subsectors of the industry and ofpentachlorophenol for three subsectors of the industry (Table A-3). However, assuminghuman consumption of both water and organisms, the agency estimated that the proposedeffluent limits would reduce, but not eliminate exceedances of the most stringent federalhealth-based ambient water quality criteria (AWQCs) for dioxin, furan, and otherchlorinated organic priority pollutants (Table A-4).28 Comparing modeled dioxin fish
27 Berry et al. (1991) noted that thorough mixing and good process controls would be essential to ensurethat no portions of the pulp are exposed to higher than the minimum chlorine charge.28 The 1993 water quality assessment of the proposed pulp and paper effluent guidelines cites the AWQCfor TCDD for consumption of water and fish as 1.3 x 10-9 µg/L (equivalent to 0.0013 ppq (EPA 1993a,Table 3-1). In its 1984 water quality criteria report for dioxin, EPA recommended ambient levels ofdioxin in the 10-5 - 10-7 cancer risk range, with 1.3 x10-9 µg/L corresponding to EPA’s estimated risk levelof 1x10-7 (EPA 1984, p. xi). Thus, the agency’s 1993 assessment based its estimates of AWQCexceedances for the proposed pulp and paper effluent guidelines on non-binding federal ambient criteria atthe lowest end of the recommended range. As noted above, however, EPA has approved binding, numericstate ambient water quality criteria for dioxin that are even less stringent than the range recommended in1984. EPA (1993a, Table 3-1) also cites an AWQC for TCDF for consumption of water and fish as 8.10 x10-8 µg/L (0.081 ppq). According to an EPA water program official, however, the agency has no official
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 19
tissue concentrations with the various advisory action levels adopted by States, EPA(1993a) estimates that the BAT proposed in 1993 would substantially reduce (by 70-95%), but not eliminate the number of State dioxin-related fish advisories in place.
Table A-3. Effluent limits (maximum for any 1 day) for existing plants usingproposed BAT process.
Subsector TCDD(ng/kkg)
TCDF(ng/kkg)
Pentachlorophenol AOX(kg/kkg)
Bleached papergrade kraft and soda
ND 359 ND 0.267
Dissolving sulfite ND 1,870 ND 3.13Dissolving kraft 300 415 ND 0.65Papergrade sulfite N/A N/A N/A 0.1ng/kkg - nanograms per metric ton (1 ng = 10-9g; 1 metric ton = 106g)kg/kkg - kilograms per metric ton (1 kg = 103g, 1 metric ton = 1000 kg, or about 2200 lbs.)ND - No detection limits of the analytical methods for TCDD and TCDF are 10 pg/L(pg = 10-12g), or 10 ppq.29
AOX - adsorbable organic halides (see discussion immediately below)Source: Fed. Reg., Vol. 58, pp. 66078-66216.
Because dioxin and other organochlorines may be toxic in trace amounts, EPAproposed to establish effluent limitations for these pollutants measured at the bleach plantwithin the mill rather than at the end of the pipe. This permits greater detection of thesepollutants before they are diluted in down-stream milling processes or wastewatertreatment. Given a large number of samples in which toxic pollutants are not detected,their estimated concentration in the pulp mill bleach plant effluent will be sensitive to howthe “no-detect” measurements are treated. (The “no-detects” signify that the actualconcentration lies somewhere between zero and the analytical detection limits.)Consistent with the agency’s standard procedures, EPA analyzed TCDD and TCDFsample data from bleach plant effluents assuming one-half detection limit values for thosecontaminants not detected in the effluent. The agency noted that a “significant portion of[the estimated] risk is associated with the use of one-half the EPA designated detectionlimit [5 pg/l] for these” pollutants (EPA 1993a). Any particular value (or point estimate)one could apply to the non-detect samples could be regarded as arbitrary. A probabilisticapproach would employ a distribution of values ranging from zero to the detection limit.
ambient water quality criterion for TCDF. It appears that the AWQC for TCDF has been inferred fromthe AWQC for TCDD on the basis of a TEF (i.e., 0.1) and a different estimated bioconcentration factor(BCF). (See EPA 1993a, p. 20. The BCF is used to estimate the concentration of a substance in fishtissue based on its concentration in water). EPA (1993a, Attachment A-12) also estimates that theproposed BAT would result in no remaining exceedances of the AWQCs for pollutants other than TCDDor TCDF if human consumption is assumed to be limited to fish and not to include drinking water.29 An 70 kg (154 lb.) person drinking 2 liters of water per day containing 10 pg/L would receive a dose of0.29 pg/kg/day. This figure is more than an order of magnitude (over forty-fold) higher than EPA’s 1985one in a million cancer risk-specific dose of 0.006 pg/kg/day.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 20
This approach is conceptually preferable to the simpler point-estimate approach, but itremains unclear how to ascertain the precise form of the distribution of pollutantconcentrations below the analytical detection limits.
Table A-4. Estimated AWQC Exceedances - Number of Streams Below PlantsExceeding AWQC by Industry Subsector
Subsector Priority Pollutant Baseline Proposed BATBleached papergrade TCDD 80 71kraft and soda TCDF 59 15
Baseline - estimated current AWQC exceedances.Proposed BAT - estimated AWQC exceedances after implementation of proposed BAT.a In addition to TCDD and TCDF, of the organochlorines listed, chloroform, pentachlorophenol,and 2,4,6-trichlorophenol are regulated as priority pollutants.Sources: Fed. Reg., Vol. 58, pp. 66078-66216; EPA (1993a, Attachment A-12).
Alternatively, all chlorinated organic compounds in an effluent sample can bemeasured collectively at the end of the pipe using an indicator such as the concentration ofadsorbable organic halides (AOX). AOX is a test measure used by Swedish researchers instudies conducted between 1977 and 1985 to evaluate some dramatic effects on fishpopulations located near bleached kraft paper mills in that country (e.g., fishkills). TheSwedish Environmental Protection Agency (SEPA) began regulatory action in 1986 toreduce the total organochlorine discharges from pulp and paper mill. In its regulation,Sweden relied on AOX because it is a relatively inexpensive and reliable measurementtechnique and because essentially all of the halides emitted from pulp and paper mills arechlorinated compounds (Thompson and Graham 1997). The SEPA established an AOXdischarge limit of 1.5-2.0 kg/air-dried ton for Swedish mills (Berry et al. 1991).
According to EPA (1993a), however, the distribution of observed effects in fishpopulations does not appear to correlate well with AOX measurements, and there is nostatistically significant relationship between the level of AOX and specific chlorinated
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 21
organic compounds, such as TCDD and TCDF. Berry et al. (1991) concluded that AOXis essentially linearly related to the amount of chlorine used in bleaching while thepolychlorinated organic materials (dioxin, furan, and polychlorinated phenols) show agreater rate of reduction than AOX as the chlorine charge is reduced. Furthermore,according to an industry official, unlike pulp and paper mills in the U.S., Swedish mills donot employ secondary wastewater treatment that can biologically degrade someorganochlorines. As a result, a Swedish mill would discharge a larger amount of organics(which deplete dissolved oxygen in water that fish require for respiration) and a differentmix of pollutants. 30 On the basis of the Five Mill Study and the integrated risk assessmentof dioxin from pulp and paper mills required under the consent decree, EPA (1993a)concludes that “although AOX concentrations can be used to determine the removal ofchlorinated organics to assess loading reductions, they do not provide information on thepotential toxicity of the effluent.”
An environmentalist argues that because discharges of dioxin andpentachlorophenol below the analytical detection limits will not meet stringent, federalhealth-based ambient water quality criteria, it is justifiable to use reductions achieved inAOX loadings as a surrogate for comparing the efficacy of alternative pollution controltechnologies. The preference for AOX is also motivated by concerns that only a few ofthe numerous organochlorines discharged by bleaching mills have been identified ortoxicologically tested and that many of the uncharacterized compounds could beenvironmentally hazardous. An industry official, on the other hand, focuses on the lack ofcorrelation between AOX and effluent toxicity and argues that if alternative pollutioncontrol technologies yield similar concentrations of dioxin and furan, it is invalid tocompare them on the basis of reductions achieved in AOX. The reductions in AOX, saysthis industry representative, do not achieve any measurable or monetizable environmentalbenefits. In addition, the estimated AWQC exceedances for dioxin are based on non-binding federal criteria that are substantially more stringent than the binding ambientcriteria established by some States and approved by EPA. According to an academic,industry has demonstrated that it can meet the dioxin AWQC of states like Marylandwithout adopting all of the measures proposed by EPA in 1993.
30 Secondary biological treatment can have a strongly reduce polychlorinated phenol discharge levels.However, the efficiency of chlorophenol removal by secondary treatment varies greatly. Secondarytreatment has a weak effect on dioxin and furan levels (Berry et al. 1991).
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 22
This seemingly arcane disagreementover the use of AOX will perhaps be thepivotal issue in finalizing EPA’s pulp andpaper effluent regulations. As would berequired under EPA’s regulatory proposals,the pulp and paper industry has already begunto convert many of its bleaching plants fromelemental chlorine to chlorine dioxide. Asdiscussed above, in some plants wherecomplete substitution of ClO2 for elementalchlorine is currently being used, the result hasbeen TCDD and TCDF concentrations ineffluents that are below the limits ofdetectability. Dioxin and furan have beendetected in other plants employing both complete substitution and oxygen delignification(AET 1994; ERG 1996). This information was unavailable at the time of the proposalbecause no pulp bleaching plants were operating using 100% chlorine dioxide prior to theproposal. (See discussion below.)
As originally proposed in 1993, the regulations would require large segments ofthe industry to employ OD (or extended delignification). Based on the new environmentalperformance data for mills employing complete substitution, the agency announced in1996 that for the bleached papergrade kraft mills, two BAT options were beingconsidered: complete substitution with and without additional delignification (Fed. Reg.,Vol. 61, pp. 36835-36858). In terms of organochlorines, the main measurable differencebetween the alternative technologies is a substantial reduction in AOX achieved by addingdelignification.31 Because there is no discernible threshold chlorine charge for dioxinformation in the pulp bleaching process, some imperceptibly small reduction in theformation of dioxin-like compounds may be associated with a reduction in AOX. Howmuch reduction there would be in concentrations below the detection limit is speculative.
31 EPA has estimated that bleached papergrade kraft mills using complete substitution and OD canachieve undetectable levels of TCDD and AOX concentrations of approximately 0.25 kg/kkg. In its 1996notice, EPA proposed to set the daily maximum limitation for bleached papergrade kraft mills usingcomplete substitution at 0.769 kg/kkg (Fed. Reg., Vol. 61, p. 36842). According to an industry official,however, using complete substitution alone yields AOX concentrations of approximately 0.5 kg/kkg.Therefore, the incremental reduction in AOX appears to be on the order of 50-70%. The reductions inAOX are also associated with reductions in conventional pollutants such as chemical and biologicaloxygen demand. NRDC (1996) reports that EPA has estimated that requiring OD in addition to completesubstitution would reduce the cumulative loading of chlorinated phenolic compounds by 2,000 kg peryear. In EPA’s 1996 notice, however, the agency indicated that both options were expected to achieveundetectable daily maximum bleach plant limits for specified chlorinated phenolics (Fed. Reg., Vol. 61, p.36841).
In terms of organochlorines, the mainmeasurable difference between requiring100% chlorine dioxide alone and requiring itin combination with oxygen delignification isa substantial reduction in AOX, a waterquality indicator that is not statisticallyrelated to levels of dioxin or any particularorganochlorine. Environmentalists justifyusing AOX because discharges of dioxinbelow the analytical detection limits will notmeet stringent federal health-based criteriaand because AOX is a surrogate for a “soup”of chlorinated organics about which little isknown. Industry counters that reductions inAOX do not achieve any measurable ormonetizable environmental benefits.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 23
3. The Process within EPA
Setting the Agenda
According to an environmentalist, the inadvertent detection of high levels of dioxinin fish downstream of pulp and paper mills in 1983 was primarily responsible for gettingthe effluent regulations on the agency’s agenda. An EPA official observes, “Even in theabsence of [detecting] dioxin, there was an internal schedule within the water program thatwould have had them [the water office] review the effluent guidelines for pulp and papermills. But the dioxin issue changed the pollutant of concern and added new impetus tothat exercise.”32 Harrison and Hoberg (1991) conclude, however, that while thepreliminary findings in 1983 were sufficient to get pulp mill dioxins quickly onto EPA’sresearch agenda, when exposure of the findings to a wider audience occurred as a resultof the 1987 New York Times article by Philip Shabecoff, pulp mill dioxins were elevated tothe regulatory agenda.33 Prominent press reports in 1989-90 concerning dioxin in milkfrom bleached paper cartons and dioxin-related fish advisories below pulp and paper millskept up the pressure (Thompson and Graham 1997).
Sources from EPA and industry agree that the Reilly administration made thedecision to combine the hazardous air pollutant and clean water rulemaking for the pulpand paper industrial “cluster” for administrative reasons, rather than as an attempt to takean integrated look at reducing environmental risks across environmental media from thesector as a whole.
Dioxin Formation Disclosed: The 5 Mill Study
Initially, EPA suspected that the source of dioxins detected in the 1983 nationalsurvey reference streams was use of dioxin-contaminated chlorophenols as “slimicides” onpulp mill machinery, rather than formation of dioxins during the process (Harrison andHoberg 1991). In 1985, the Environmental Defense Fund and the National WildlifeFederation filed suit against EPA for denying the environmentalists’ petition to regulatedioxins and furans under TSCA. Meanwhile, EPA tested wastewater treatment sludgefrom pulp and paper mills and found that dioxin levels were highest in the sludges ofbleached kraft pulp mills. This suggested that dioxin was probably being formed as a by-product during the bleaching of wood pulp with chlorine. (Fed. Reg., Vol. 58, p. 66092).In 1986, EPA, the American Paper Institute (API), and NCASI (National Council of thePaper Industry for Air and Stream Improvement, the industry’s research arm) agreed toundertake the “5 Mills Study.” The study results detected TCDD and TCDF in effluents
32 CWA Sec. 307 (a)(3) requires effluent standards to be reviewed every three years.33 Shabecoff and his successor at the NYT, Keith Schneider, have been prominent figures in the recentdebate over the appropriate role of environmental journalists. Many viewed Shabecoff and otherenvironmental journalists as being too sympathetic to environmental groups, and Schneider, who wrote ahighly-publicized series of articles on EPA’s over-estimation of the cancer risks of dioxin, went on tobecome the leading voice of a revisionist camp of environmental journalism.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 24
of four of five mills, pulps of all five mills, and wastewater treatment plant sludges of allfive mills.
Environmentalists learned of the agreement between EPA and the paper industryto conduct the 5 Mill Study, and in December of 1986, Greenpeace initiated a Freedom ofInformation Act (FOIA) request seeking all available information on the pulp mill dioxinproblem. In January of 1987, a letter from an EPA official to API was leaked whichsuggested that EPA had agreed to notify the industry immediately of receipt of any FOIArequests and that, barring such requests or results indicating a potential threat to humanhealth, the agency did not intend to release any results until publication of the final 5 MillsStudy report. The following August, Greenpeace released a report alleging an EPAcover-up (Harrison and Hoberg 1991). In September, based on the 5 Mills Study andanalyses of dioxin in paper products that were later added to the study, the New YorkTimes ran a front-page story reporting that traces of dioxin had been detected inhousehold paper products (Shabecoff 1987). Until this point, dioxin was primarilyassociated in the public mind with pesticides and combustion processes. The prominentdisclosure that dioxin was formed by pulp bleaching and was present in commonhousehold products undoubtedly provided EPA with leverage to gain industry’scooperation to support additional research.
Integrated Assessment, the Consent Agreement, and the 104 Mill Study
According to an EPA official, when the agency’s water program first realized thatdioxin was being formed by pulp and paper bleaching, it immediately recognized that theproblem went beyond its jurisdiction and approached the toxic substances program saying,“we have a problem to share with you.” After a quick review, the agency decided that theissue spilled over into the risk management jurisdictions of other agencies. An interagencyworkgroup including EPA, FDA, the Consumer Products Safety Commission (CPSC),and the Occupational Safety and Health Administration (OSHA) was formed to identifythe data needs to determine the extent of the problem. The interagency group met withAPI to lay out its plans for a multi-media, multi-pathway assessment of effluents, pulp,sludge, occupational, and consumer risks and requesting industry to bear the burden of thecosts. An industry official says that EPA proposed using its TSCA (Sec. 4) authority torequire the pulp and paper sector to provide the data, and noted that the authority suppliedthe agency “with an arrow in their quiver” during the negotiations. The American Forestand Paper Association (AF&PA, a newly consolidated industry trade association) “hit theroof” because of the TSCA threat, says this source. “It was a question of trust.” EPA didnot exercise its TSCA authority but retained it as a negotiating point. Thus began the“104 Mill Study.” The assessment officially began in 1989, and an EPA official noted thatthe forest products industry spent $3-$4 million on the study.
According to an EPA official, to get a handle on the extent and magnitude of theproblem, the interagency group insisted on representative samples from all 104 mills thatused chlorine in bleaching. EPA assumed the coordinating role and concentrated on therisks from effluents, sludge, and occupational exposures. The EPA coordinator was a
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 25
Dwain Winters, an analyst in the Office of Toxic Substances (OTS). The FDA focused onfood contact papers and medical devices. The CPSC looked at writing papers, diapers,and other consumer products. Overall, the group identified more than 153 separate dioxinexposure pathways to be analyzed. The interagency group developed a QualityAssurance/Quality Control protocol for conducting samples, and NCASI conducted thestudy.
In securing broad coverage to provide a strong analysis for the purposes ofscoping and identifying the dioxin problem from bleaching mills, the interagency grouptraded off in-depth analysis at individual plants that would have been more useful informulating a technological remedy. An academic observes that this investigative strategyalso permitted the agency to avoid having to generalize between different mills usingsimilar processes. Industry, on the other hand, wanted to do an intensive study ofparticular types of bleaching mills in order to better evaluate processes which causeddioxin formation and to identify the key steps that were responsible. Late in 1988, theindustry designed and conducted an intensive study of 22 bleaching plants independent ofthe 104 Mill Study (Thompson and Graham 1997).
While the negotiations with industry were ongoing, the agency began separatenegotiations with EDF and NWF regarding their suit by describing the planned study. Inaddition to writing the 104 Mill Study into the consent decree, EPA agreed to makedeterminations whether regulatory actions were required for the pulp, sludge, and effluentand, if so, to identify the information needed for regulatory decisionmaking. Thesedecisions were subject to judicial review. An academic notes that by involving EDF andNWF in negotiating the plan for the 104 Mill Study, the scope of the study was broadenedand the process was somewhat delayed.
According to an EPA official, as a result of the 104 Mill Study, dioxin in pulp andpaper wastewater was identified as the route of major health concern, with land disposalof sludge being a secondary health concern. Agency analysts, however, flagged landdisposal of sludge as the primary route of ecological concern. The agency did not viewdioxin in paper products as a major risk, but it was “on the borderline of concern,” so theagency referred food contact papers to FDA under TSCA (Sec. 9). FDA accepted thereferral and pursued voluntary reductions, according to this source. To arrive at theseconclusions, the agency scientists combined the results of the 104 Mill Study withinformation regarding the health hazards of dioxin and analyses of various exposurepathways. Some of this information was taken as given (e.g., the estimated carcinogenicpotency of dioxin supplied by the 1985 EPA hazard assessment), some of it representeddepartures from EPA’s normal assessment procedures for water quality criteria (e.g.,assumptions made regarding fish consumption), and some of it resulted from originalresearch (e.g., a study was conducted on uptake of dioxin through the skin using cadavertissue samples).
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 26
Hazard Assessment
EPA had established its official position on the environmental and health effects ofdioxin before the integrated assessment got underway. The agency issued its AmbientWater Quality Criteria Document for dioxin in 1984.34 According to the hazardassessment issued by EPA in 1985, dioxin was to be regarded by all agency programs as aprobable human carcinogen on the basis of adequate animal data and limited human data.The dose associated with an increased cancer risk of up to 10-6 was officially 0.006 pg/kgbody weight/day. The agency’s 1983 statement permitting states to use “otherscientifically defensible methods” to modify EPA ambient water quality criteria, as well asthe agency’s initial stab at a dioxin reassessment in 1988, invited the States to exercisesome scientific discretion in assessing the hazards of dioxin. For the EPA regulatoryprograms offices, however, using the results of the 1985 assessment remains non-discretionary. According to an EPA water official, it is clear that the toxicologicalinformation on dioxin available through the agency’s Integrated Risk Information System(IRIS) was to be used in the program’s analyses, “we use whatever ORD (the agency’sOffice of Research and Development) tells us to use.”
The latest dioxin reassessment was underway by 1991, and the water program hasacknowledged in its regulatory proposals that new information might become available,but there seems little chance that the agency will reach closure on the review of dioxin’stoxicity before it finalizes the pulp and paper effluent guidelines. According to an EPAofficial, however, the minor adjustment made to the dioxin risk-specific factor in the 1994draft reassessment (from 0.006 pg/kg/day to 0.01 pg/kg/day for a cancer risk of 10-6)would not have significantly affected decisions about the pulp and paper sector. Theexisting Reference Dose (RfD) on the IRIS for dioxin non-cancer effects is at about thelevel where cancer risks approach 10-4 (one in ten thousand) Because the agency wasconcerned with individual risk levels below this (10-5 - 10-6), says this source, estimatedcancer effects drove the determination that the pulp and paper effluent was the primehuman health concern. An EPA water official comments, “At this stage, on the effects ofdioxin, as a user of science, I feel somewhat more certain. But for every question we’veanswered, we’ve raised new ones.”
The scope of EPA’s hazard assessment was limited by the complex and variablechemical composition of pulp and paper effluents and by the lack of toxicity data for thosesubstances which were identified. Based on an evaluation of pulp and paper effluentsampling data collected by EPA (both independently and in cooperation with industry), theagency identified 26 organic chemicals (including dioxin and furan) as contaminants ofconcern. Of these 26 contaminants, 24 are organochlorines and 6 are priority pollutants.Only 11 have RFDs and 6 have cancer potency factors available using EPA’s primarytoxicological databases, IRIS and HEAST (Health Effects Assessment Summary
34 The 1984 Criteria Document was prepared jointly by the EPA Offices of Water and Research andDevelopment (ORD). The process was managed by the ORD Environmental Criteria and AssessmentOffice located in Cincinnati, OH. The health effects chapter acknowledged 44 contributors from EPA andother federal agencies, international institutions, academia, industry, and environmental groups.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 27
Tables).35 Due to a lack of data on human health toxicity, only 13 of the 26 contaminantscould be evaluated for their potential human health impacts (EPA 1993a).
Exposure Assessment
The dioxin exposure assessment was conducted by a consultant (Tetra Tech) undercontract to the EPA Office of Water/Office of Science and Technology/Standards andApplied Science Division. To conduct the assessment required estimates of severalfactors. The factors and the sources of the estimates are provided in Table A-5. Anexamination of the table reveals a grab bag of types and sources of scientific information:peer reviewed literature, gray literature, default assumptions, ad hoc assumptions, andprofessional judgments produced by all levels of government, industry, and academia.
Table A-5. Sources of the Estimates Used in Dioxin Exposure Assessment.
Exposure Factor Source of EstimateDioxin discharges from plants 104 Mill StudyIn-stream dioxinconcentrations
Site-specific flow data for 68 streams, dilution factors for 17 millsdischarging to open waters (e.g., oceans, estuaries, lakes), and twodilution models. Dilution factors provided by Office of Waterpublication and Regional EPA personnel. One of the dilution modelswas developed by EPA/Office of Research and Development/ExposureAssessment Group and was still under EPA review at the time.
Dioxin concentrations in fish Bioconcentration factor (BCF) for TCDD in trout from a laboratorystudy included in 1991 Banbury Report edited by Gallo et al.BCF for TCDF derived from 1988 article in EnvironmentalToxicology and Chemistry. Biota to Suspended Solids AccumulationFactor (BSSAF) of 0.09 derived from EPA Lake Ontario study.BSSAF of 0.02 derived from NCASI study.
Fish consumption rates For recreational anglers, estimates based on a 1988 New York Statesurvey; a 1981 Tacoma, WA County Health Department Report; and a1989 University of Michigan Technical Report.For subsistence anglers, 145 g/day assumed and said to be“consistent” with a 1982 U.S. Dept. of Ag. report.
Change in fish consumptiondue to State fishing advisories
EPA assumed a 20% decrease for recreational anglers and no effect onsubsistence anglers based on a 1990 University of Michigan M.S.Thesis and a 1990 paper presented at an American Fisheries SocietyMeeting.
Size of fish-consumingpopulations
Number of fishing licenses in a county multiplied by average familysize (based on Census Bureau data) of 2.63 to yield fish-consumingpopulation in each county. EPA assumed that 95% of fishing licenseswere issued to recreational anglers and 5% to subsistence anglers.
Exposure duration, bodyweight, etc.
EPA Office of Solid Waste and Emergency Response StandardDefault Exposure Factors
35 The priority pollutants are chloroform, methylene chloride, pentachlorophenol, TCDD, TCDF, and2,4,6-trichlorophenol.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 28
The exposure assessment for dioxin in pulp and paper effluents marked a departurefrom EPA’s standard operating procedures for developing water quality criteria.Normally, the water office would assume fish consumption levels based on nationalaverages. In this case, however, the agency focused on exposure scenarios it expectedwould occur near pulp and paper mills, including highly exposed subpopulations--sportsanglers and subsistence anglers. According to an EPA official, for the fish consumptionrates of highly exposed subpopulations, risk analysts can only provide some reasonableupper-bound estimate that is subject to a high degree of uncertainty due to the lack ofdata. In contrast, there are good data available on average fish consumption rates by thegeneral population. This source allows that “the way we chose to calculate the risk forfish consumption became complicated. The normal procedure under water quality criteriaassumed a certain level of average consumption, but we had to make assumptions thatwent beyond that to look at specific subpopulations like subsistence anglers. Nobodybelieved that the average consumption rates would be representative of real exposures, sothe solution was to provide a number of consumption rates.”
EPA also evaluated the exposure scenarios under two alternative models forestimating how dioxin concentrations would be diluted in receiving water bodies. Thefirst was a simple dilution model, which assumes that all carcinogenic pollutantsdischarged into a receiving stream are available for uptake by fish. The other--called theDRE model--is a more complex dilution model under development by EPA . It assumesthat the uptake of dioxins and furans depends on the levels of suspended solids and thepartitioning of the pollutants between fish tissue and sediment. As a result, thepartitioning model produces lower exposure estimates.
In EPA’s 1996 notice that it was considering two BAT options for the major pulpand paper subsector, the agency also stated that it was considering using only the DREmodel for estimating dioxin and furan concentrations in fish for the final rule. However,the agency stated that it would modify the DRE model, however, to reflect ongoingcontamination. The modification entailed replacing the Biota to Suspended SolidsAccumulation Factor (BSSAF) of 0.09 with a BSSAF factor of 0.2 (Fed. Reg., Vol. 61, p.36846). This might seem trivial, but the BSSAF is a critical model parameter in predictingthe accumulation of dioxin in fish. The BSSAF value of 0.09 represents the agency’sdefault for dioxin based on Lake Ontario data which is primarily from historical sources.To support its conclusion that a BSSAF factor of 0.2 was more appropriate forecosystems subject to ongoing contamination, the agency cited EPA (1994c), one of thevolumes of the agency’s draft dioxin reassessment. Although EPA (1994c) indicates thatBSSAFs for aquatic systems where contamination is ongoing might be greater than fromsystems where the contamination is primarily historical, it stops short of endorsing anyparticular value for all water bodies where contamination is ongoing. EPA (1994, Sec.7.2.3.6) describes an analysis of the impact of pulp and paper mill effluents on fish tissueconcentrations conducted by NCASI using data from the 104-mill study. This study foundthat for the 38 mills discharging into smaller water bodies, using a BSSAF of 0.2 (upfrom 0.09) improved the predictive performance of the model. According to EPA (1994c,Table 4-1), BSSAFs for dioxin observed in the literature range from 0.009 to 2.94.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 29
Therefore, EPA’s assumption that a BSSAF of 0.2 is appropriate for all waterbodies downstream of discharging mills, regardless of the size of the receiving water bodyand other factors, appears to have a somewhat tenuous basis. More importantly, however,this analytic anecdote illustrates the problems associated with attempting to conduct anenvironmental risk assessment for a national rulemaking while taking into account site-specific conditions. On the basis of at least some relevant, empirical data, EPA isadjusting its default model to take into consideration differences between historical andongoing dioxin contamination. Making a finer distinction could potentially require a largenumber of field studies to determine the BSSAF for every water body downstream of ableaching pulp and paper mill. Furthermore, despite the importance of the BSSAF inpredicting the dioxin accumulation in fish tissue, it is only one of many exposure factors tobe evaluated.
The Regulatory Impact Assessment: The Numbers Behind the Proposal
The Water Office’s Regulatory Impact Assessment (RIA) estimated that theproposed BAT for the pulp and paper industry would result in a national annual reductionof 5-35 cancers in recreational and subsistence anglers (EPA 1993b). The lower value inthe range is only for TCDD and TCDF and was estimated using the partitioning dilution(DRE) model. The upper value is calculated using the simpler, conservative dilutionmodel and is still dominated (99%) by the effects of TCDD and TCDF. Using a range of$2 to $10 million for the value of a life, EPA (1993b) estimated the benefits of theproposed effluent regulations to be $10 - $350 million per year.36 Using the simpledilution approach, EPA (1993a) estimated that greater than 99 percent of the noncancerhazard (as indicated by the mills exceeding RfDs for recreational and subsistence anglerpopulations) can be attributed to dioxin and furan.
Senior EPA sources have suggested that the industry has framed the issuenarrowly by focusing on cancer effects and failing to consider non-cancer effects ofdioxins. An EPA water official, however, notes that the program was unable to monetizethe benefits of non-cancer health effects in the Regulatory Impact Assessment (RIA) dueto the lack of quantitative dose-response data for non-cancer effects. (This situation isgeneric to non-cancer effects and not unique to dioxins.) According to this official, “wedid a very good job with what science we had.” Not surprisingly, an industry officialcharacterizes the RIA as a “poor job.” An environmentalist complains that the Office ofWater feels compelled to conduct a detailed risk and economic assessment for a“technology-based” rule. However, EPA is required by Executive Order to conduct aRegulatory Impact Analysis for all proposed regulations with an annual economic impactof more than $100 million.
36 According to Viscusi (1996), on average, workers receive $3 - $7 million in compensation per statisticallife lost.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 30
Communicating the Science to Decisionmakers
Because Assistant Administrator for Water Robert Perciasepe was confirmed onlydays before the 1993 proposed effluent regulations were issued in compliance with theconsent decree, he was unable to engage in the decisionmaking. According to an EPAwater official, the key decisionmakers included Acting Assistant Administrator MarthaProthro and Tudor Davies, Director of the Water Office of Science and Technology (thenacting Deputy Asst. Adm.). Reportedly, EPA Administrator Carol Browner was alsobriefed. The issue was of sufficient importance that it was not completely delegated to anActing Assistant Administrator (AA). “The Cluster Rule was high stakes, high visibility,”says this source, adding “this involved the ‘D-word’ [dioxin]. It had everybody’sattention.” There were presentations for Prothro from NCASI, AF&PA, and academics.“The AA attended mill [plant site] visits--that’s unusual.” Due to the high level ofdecisionmaker interest, according to this source, it was “not difficult to get theirattention.” In addition, because dioxin is regarded as an important multi-media pollutant,the decisionmakers “knew a lot about it; they were sensitized and could mesh thingstogether about dioxin other than what they just heard about in the context of pulp andpaper.”
4. The Proposal, Industry’s Response, and a New Framework
In announcing the 1993 proposal, Assistant Administrator Perciasepe estimatedthat dioxin discharges from the pulp and paper sector into surface waters would bereduced from over 300 grams to less than 30 grams per year (Environment Reporter,11/5/93, pp. 1227-1228). The BAT effluent regulations proposed in 1993 for four pulpand paper subsectors organized by production process are summarized in Table A-6.
In 1993, NCASI reported data on the pulp and paper industry’s progress inreducing the dioxin content of effluents, pulps, and wastewater treatment sludges (NCASI1993). In 1994, a study prepared for a supplier of chlorine dioxide technology, theAlliance for Environmental Technology (AET), reported that substitution of chlorinedioxide for elemental chlorine reduces the formation of chemicals like dioxin thataccumulate in fatty tissues to the limits of detectability (AET 1994). Among the study’sauthors was Robert Huggett, a marine biologist at William and Mary College, and soon-to-be-EPA Assistant Administrator for ORD.37
37 Huggett recused himself from involvement in the pulp and paper dioxin rule.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 31
Table A-6. 1993 Proposed BAT and Estimated Impact for Pulp and PaperSubsectors.
Subsector Principal Products Proposed BAT Size/Est. CostBleached papergradekraft and soda
papergrade kraft marketpulp, paperboard, coarsepapers, tissue papers, andfine papers for business,writing, and printing
OD* and 100% ClO2 78 mills$260 annualized1-3 plant closures500-4,000 jobs
Dissolving sulfite pulps used for rayon,cellophane, and celluloseproducts
OD and 100% ClO2 5 mills$5 annualized1 plant closure
Dissolving kraft pulps used for rayon,acetate, and other celluloseproducts
OD and 70% ClO2 3 mills$11.9 annualizedNo plant closures
Papergrade sulfite tissue paper, fine papers,newsprint
TCF 10 mills$25 annualized2 plant closures
OD* - oxygen delignification or extended cooking of pulp prior to bleaching.OD - oxygen delignification prior to pulp bleaching.100% ClO2 - complete substitution of elemental chlorine (Cl2) for chlorine dioxide (ClO2) in thebleaching process.70% ClO2 - seventy percent substitution of Cl2 for ClO2.TCF - Totally Chlorine Free bleaching using peroxide (H2O2) or ozone (O3).Costs in millions of 1991 dollars for all the mills in the subsector, as estimated by EPA.38
The new data on the performance of alternative technologies made it clear to EPAsources that human health risks resulting from land disposal of pulp and paper sludgewere moot. Agency officials felt confident that whatever the final outcome of the ClusterRule, it would simultaneously control human health risks associated with sludge landdisposal. EPA (1990) estimated that TCDD soil concentrations of 0.12 ppt and 12.0 pptwould produce human cancer risks of 10-6 and 10-4, respectively. The new performancedata suggested that alternative technologies achieved overall decreases in the formation ofdioxins. This meant that the dioxins were not simply being shifted from the wastewater tothe sludge.
Some EPA official, however, remained concerned about possible ecologicaleffects of sludge land disposal. EPA (1990) estimated that if sludge were to be landapplied, dioxin concentrations as low as 0.03 ppt would be necessary to protect the “mostsensitive” terrestrial species, identified as the American woodcock (Scolopax minor). Thebird would be highly exposed to dioxin in soils because the contaminant accumulates in
38 According to an industry official, EPA underestimated the capital costs of the proposed BAT by morethan 40% because OD, in combination with some of the best management practices (BMPs) EPA hasproposed (particularly spill controls) may require plants to install larger recovery furnaces (costingapproximately $100 million per plant).
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 32
the woodcock’s primary food source, worms.39 Some EPA staff felt that the agency’sdecision to retract the proposed regulation of sludge land disposal did not give dueconsideration to ecological risks. They also felt that environmental advocates who hadbeen engaged in the issue lost interest once human health risks appeared under control.According to an industry official, however, the AF&PA insisted upon an SAB review ofthe ecological risk assessment. The SAB determined that while the agency had identified ahazard to the woodcock, it had not substantiated that population-level effects would resultfrom land disposal of dioxin-tainted pulp and paper sludge.40 In 1994, EPA negotiated avoluntary agreement with the pulp and paper industry (Gilman and EPA 1994) that placesno land disposal restrictions on sludges containing less than 10 ppt dioxin and servesprimarily to formalize best management practices.
In 1995, EPA announced the availability of sampling data characterizing theperformance of bleached papergrade kraft and papergrade sulfite mills employingcomplete substitution of elemental chlorine with ClO2 with and without oxygendelignification (Fed. Reg., Vol. 60, pp. 34938-24940). In July 1996, EPA announced thatits review of new data on the performance of chlorine dioxide bleaching indicated thatdioxins and furans in wastewater discharges from bleached papergrade kraft and soda millscould be reduced by 95 percent and 99 percent, respectively (EPA 1996). The agency alsoannounced a “new framework” for the pulp and paper effluent guidelines. The primaryelements of this new regulatory framework included: a) consideration of two BAToptions for two pulp and paper subsectors: bleached papergrade kraft and papergradesulfite; b) deferral of a final decision for the dissolving sulfite and dissolving kraftsubsectors; and c) a voluntary incentives program to reward mills that exceeded regulatoryrequirements in reducing discharges (Fed. Reg., Vol. 61, pp. 36835-36858).
For the bleached papergrade kraft subsector, EPA announced that it wasconsidering complete (100%) substitution of ClO2 for elemental chlorine (Option A) andcomplete substitution with OD (or extended delignification) (Option B). EPA stated thatboth options appear to reduce dioxins and furans in wastewaters to concentrations at orbelow the current analytical detection limits. The incremental environmental benefits thatEPA attributed to the use of OD (or extended delignification) included reduced chronictoxicity to some aquatic species. EPA stated that the reduced chronic toxicity is probablyattributable to a reduction in chemical oxygen demand. (COD is a conventional waterpollutant, the technological control of which is subject to a test of economicreasonableness, as discussed above.) The agency also stated that the reduced chronictoxicity may reflect an incremental reduction in the potential formation of dioxin and furan 39 The woodcock is a favored game bird. Adults woodcocks are approximately 1 foot long and weigh halfa pound and can eat their weight in worms daily. In response to the drop in woodcock numbers over thelast 20 years in some parts of its range (from the Georgia to the Canadian maritime provinces and west toMinnesota), the U.S. Fish and Wildlife Service announced the North American Woodcock ManagementPlan in 1990. Because the woodcock’s breeding habitat is successional scrub forest, the principal cause oftheir decline appears to be the maturation of hardwood forests growing on abandoned farmlands in thenortheast (http://alloutdoors.com; //www.im.nbs.gov).40 It should be noted that estimating population-level effects is extremely uncertain and can requirelengthy and costly site-specific studies.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 33
below the analytical detection limits as well as a reduction in all chlorinated compoundsloadings, as measured at the end-of-the-pipe by AOX. (Recall that dioxin and furan arepriority pollutants, the technological control of which is subject to the less demandingeconomically achievable standard.) The notice further states:
Although statistically significant relationships between AOX and a broadrange of specific chlorinated organic compounds have not been established,trends in concentrations changes have, however [sic], been observedbetween AOX and specific pollutants, including dioxin, furan, andchlorinated phenolic compounds. Even though dioxin and furan are nolonger measurable at the end-of-pipe at many mills, the potential forformation of these pollutants continues to exist at pulp and paper mills aslong as any chlorine-containing compounds (including chlorine dioxide) areused in the bleaching process...EPA expects that [reductions in] AOXdischarges...will in turn further reduce the likelihood of the formation anddischarge of these chlorinated organic pollutants.
Thus, EPA appears to reject the claim that there is a threshold level of chlorine belowwhich no dioxin formation occurs in bleaching pulp. The agency also implicitly suggeststhat it is considering requiring reductions in priority pollutants beyond the point ofdiminishing returns because such reductions are achievable. (Recall that thedetermination of BAT involves consideration of both technological and economicachievability.)
Table A-7 summarizes the proposed effluent limits under the 1996 Options A andB for the Bleached Papergrade Kraft and Soda subsector. Table A-8 summarizes EPA’sestimates of the economic impacts under Options A and B (Fed. Reg., Vol. 61, pp. 36840-36841).
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 34
Table A-7. Bleached Papergrade Kraft and Soda Plant (Daily Maximum) Limitations
OD* - oxygen delignification or extended cooking of pulp prior to bleaching.ND - No detect.
Table A-8. Estimated Impacts for the Bleached Papergrade Kraft and Soda Subsectorunder 1996 Options A and B.44
Option B 1993OD* and 100% ClO2
Option A 1996100% ClO2
Option B 1996OD* and 100% ClO2
78 mills$223.2 Annualized1-3 plant closures500-4,000 jobs
85 mills$140 Annualized1 plant closure500 jobs
85 mills$155 Annualized3 plant closures4,100 jobs
OD* - oxygen delignification or extended cooking of pulp prior to bleaching.Costs in millions of 1995 dollars, as estimated by EPA.
5. Concluding Observations
Sources interviewed for this case study suggest that the treatment of AOX is theprincipal outstanding science policy issue regarding the final BAT determination for thepulp and paper effluent guidelines. In terms of a fate and transport analogy, theinformation provided by AOX was first transported to EPA from Sweden in the early1980s via international scientific journals. In 1986, the use of AOX as a regulatoryparameter achieved a measure of international legitimacy when the Swedish EPA adoptedit. Australia, Austria, Belgium, Finland, and Germany have followed, and AOX wasthoroughly assimilated by EPA into the 1993 proposal. However, the fate of theinformation provided by AOX varied considerably across countries. The AOX limitscurrently being considered by EPA (0.448 kg/ton or .162 kg/ton based on a monthly
41 The difference in units for TCDF between the 1993 proposal and the 1996 options reflects a changefrom a production-normalized bleach plant limitation to an effluent concentration-based limitation.42 ND for two pollutants (trichlorosyringol and 2,4,6-trichlorophenol)(mg/kkg).43 Whereas the other pollutant limits are measured at the bleach plant, that for AOX is an end-of-pipemeasurement.44 Differences in estimated impacts between the 1993 proposal and Option B in 1996 in Table A-7 reflecta revised economic impact analysis. The difference between Tables A-6 and A-7 in the annualized costestimate for the 1993 proposal is due to expressing the costs in different base year dollars (1991 or 1995).As indicated above, industry sources believe that EPA has substantially underestimated the incrementalcapital costs of OD.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 35
average limitation) are slightly to 3 times lower Sweden’s, currently the strictest national-level standard for softwood pulp mills (EKA 1996; Thompson and Graham 1997).45 Thefate and transport of AOX in international pulp and paper effluent limits is illustrated inFigure A-1.
Figure A-1. Fate and Transport of AOX in International Pulp and Paper Effluent Limits(kg/ton for softwood pulp mills).
*There is no national Canadian AOX effluent limit. It varies by province.
Although the formation of dioxin and other organochlorines from pulp bleachingseems plausible at even minimal levels of chlorine input, the effect that AOX has on EPA’scomparison of dioxin control alternatives is impeded by the agency’s inability to transformAOX into toxicological equivalent units. According to one EPA official, “on AOX, wehave a lot more information, and some specific questions have been addressed, but I don’tthink we’ve moved too far in being more certain about its environmental significance.While the uncertainty has gone down some, the level of controversy may have heightened.As the analysis now stands, we would continue to use AOX as a regulatory parameter.”This is a case, add this source, where perfect information might make the decision for thedecisionmaker.
An environmentalist suggests that the impact of science has been very high--butnegative--in this case. The Clean Water Act “is a technology-based statute, and EPA has
45 Softwoods are evergreens. Hardwoods are deciduous. In the US, most pulp is produced from softwood.Sweden has set an AOX limit of 0.3 kg/ton for hardwood pulp mills to be achieved in 2000-2005 (EKA1996).
SwedishScientists
ScientificJournals
AOX
AustraliaFinland Belgium Australia Sweden US Canada*
Norway Austria
Germany
AOXEffluentLimits
FinlandFinland
2.0 1.5 1.0 0.5 0.162-0.448
1.5-0
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 36
digressed from the technology basis.” This source points out that requiring OD inaddition to complete substitution is superior in preventing organochlorine discharges intothe environment and complains that the “health effects analysis has tied [EPA] up inknots.” This source blames industry and OMB for pressuring EPA to emphasize cost-effectiveness criteria and to stray too far from the congressional intent to rely ontechnology-based standards to achieve environmental results in the face of scientificuncertainty. This source also points out that the Natural Resources Defense Council stillhas a 1993 petition pending with EPA to ban all discharges of dioxin from the pulp andpaper industry under CWA Sec. 307a. The NRDC’s position is that due to thetechnological inability to detect dioxins at concentrations believed to be toxic, the ban canonly be achieved by prohibiting the use of chlorine in pulp bleaching.
An EPA official appreciates that when one evaluates a technology-based standardin terms of its adequacy to protect health and safety, it may be seen as falling short of themark. “For some people, any exposure to dioxin is unacceptable.” The pulp and paperindustry, once widely perceived as environmentally recalcitrant, has taken some voluntaryand negotiated steps to demonstrate how it could dramatically reduce dioxin discharges.These steps were taken in an attempt to get out ahead of the regulatory curve and inresponse to consumer demand for environmentally-friendly products and potential tortliability concerns. 46 The industry may have also sensed that the ongoing dioxinreassessment would not produce substantial regulatory relief.
In February 1997, Assistant Administrator Perciasepe announced that the pulp andpaper effluent rulemaking would not be released for several months (EnvironmentReporter, 2/12/97, p. 2142). An April 1997 article appearing in The Washington Postsuggested that the “Pulp Friction” saga had spilled over into the halls of Congress(Skrzycki 1997). In light of the performance of newly adopted pulp bleachingtechnologies, it remains to be seen whether EPA will judge the substantial and measurablereductions in discharges of dioxins and other organochlorines as sufficient to scale backfrom its 1993 proposal to impose more stringent and costly pollution control technologies.To a considerable extent, the decision hinges on the marginal “unmonetizable” benefits oftechnologically imperceptible reductions in the formation of dioxin, which, in the eyes ofsome, remains the Darth Vader of chemicals.
46 See Thompson and Graham (1996) for further discussion of the possible role of tort liability risk in thepulp and paper industry’s decision to adopt advanced bleaching technologies.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 37
REFERENCES
AET (The Alliance for Environmental Technology). 1994. A Review and Assessment of the EcologicalRisks Associated with the Use of Chlorine Dioxide for the Bleaching of Pulp.
AET (The Alliance for Environmental Technology). 1995. Five Great Reasons Why We Care: The Pulpand Paper Industry's Virtual Elimination Strategy. September. (http://aet.org/science/5reasons.html.)
Berry, R., C. Luthe, R. Voss, et al. 1991. “The effects of recent changes in beached softwood kraft milltechnology on organochlorine emissions: An international perspective,” Pulp and Paper Canada, Vol.92, No. 6, pp. 43-55.
Colborn, T., D. Dumanoski, and J. Peterson Myers. 1996. Our Stolen Future: Are We Threatening OurFertility, Intelligence, and Survival? Dutton: NY.
Copeland, C. 1993. Toxic Pollutants and the Clean Water Act: Current Issues. Congressional ResearchService Report for Congress (93-849 ENR).
CRS (Congressional Research Service). 1993. Toxic Pollutants and the Clean Water Act: Current Issues.CRS Report 93-849 ENR.
EKA (EKA Chemicals, Inc.). 1996. EKA Chemicals Comments on Cluster Rule Proposal. August 12,1996. EKA Chemicals, Inc., Marietta, GA.
EPA (U.S. Environmental Protection Agency). 1989. Interim Procedures for Estimating RisksAssociated with Exposures to Mixtures of Chlorinated dibenzo-p-dioxins and dibenzofurans (CDDs andCDFs) and 1989 Update. Risk Assessment Forum.
EPA (U.S. Environmental Protection Agency). 1990. Assessment of Risks from Exposure of Humans,Terrestrial and Avian Wildlife, and Aquatic Life to Dioxins and Furans from Disposal and Use of Sludgefrom Bleached Kraft and Sulfite Pulp and Paper Mills. Office of Pesticides and Toxic Substances, July.
EPA (U.S. Environmental Protection Agency). 1993a. Water Quality Assessment of Proposed EffluentGuidelines for the Pulp, Paper, and Paperboard Industry. Office of Water (EPA-821-R-93-022).
EPA (U.S. Environmental Protection Agency). 1993b. Regulatory Impact Assessment of proposedEffluent Guidelines and NESHAP for the Pulp, Paper, and Paperboard Industry. Office of Water.
EPA (U.S. Environmental Protection Agency). 1994a. Estimating Exposure to Dioxin-Like Compounds.Volume 1: Executive Summary. External Review Draft. Office of Research and Development.
EPA (U.S. Environmental Protection Agency). 1994b. Health Assessment Document for 2,3,7,8-Tetrachlorodibenzo-p-Dioxin (TCDD) and Related Compounds. Vol. III of III. External Review Draft.
EPA (U.S. Environmental Protection Agency) 1996. EPA releases new data on pulp and paperdischarges; considers regulatory and voluntary options. July 3 press release.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 38
EPA/SAB (EPA/Science Advisory Board). 1995. A Second Look at Dioxin.
ERG (Eastern Research Group, Inc.). 1996. Pulp and Paper Mill Data Available for BAT LimitationsDevelopment. Prepared under contract for USEPA Office of Water. July 10.
Fingerhut, M., W. Halperin, B. Marlow, et al. 1991. “Cancer Mortality in Workers Exposed to 2,3,7,8-Tetrachlorodibenzo-p-Dioxin,” New England Journal of Medicine, Vol. 342, pp. 212-218.
Finkel, A. 1988. “Dioxin: Are We Safer Now than Before?,” Risk Analysis, Vol. 8, No. 2, pp. 161-165.
Fogarty, J. 1991. “A Short History of Federal Water Pollution Control Law,” Clean Water Deskbook,2nd ed. Environmental Law Institute: Wash., DC, pp. 5-20.
Gilman and EPA. 1994. Memorandum of Understanding between Gilman Paper Company and the U.S.Environmental Protection Agency Regarding the Land Application of Pulp and Paper Mill Materials.
Hanmer, R. 1988. Memorandum from Rebecca W. Hanmer, Acting Asst. Adm. for Water to WaterManagement Division Directors, “Interim Strategy for the Regulation of Pulp and Paper Mill DioxinDischarges to the Waters of the United States, August 9.
Harrison, K. and G. Hoberg. 1991. “Setting the Environmental Agenda in Canada and the United States:Canadian Journal of Political Science, March, pp. 3-27.
Hirshfield, A., M. Hirshfield, and J. Flaws. 1996. “Problems Beyond Pesticides,” Science, Vol. 272, pp.1444-1445.
Jasanoff, S. 1990. The Fifth Branch: Science Advisors as Policymakers. Harvard Univ. Press:Cambridge, MA.
Kociba, R. D. Keyes, J. Beyer, et al. 1978. “Results of a Two Year Chronic Toxicity and OncogenicityStudy of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin in Rats,” Toxicology and Applied Pharmacology, Vol. 46,pp. 279-303.
Lucier, G., G. Clark, C. Hiremath, et al. 1993. “Carcinogenicity of TCDD in Animals,” Toxicology andIndustrial Health, Vol. 9, pp. 631-668.
Moore, J. R. Kimbrough, and M. Gough. 1993. “The Dioxin TCDD: A Selective Study of Science andin M. Uman, ed. Keeping Pace with Science and Engineering. National Academy of
Engineering: Wash., DC, pp. 221-242.
NCASI (National Council of the Paper Industry for Air and Stream Improvement). 1993. Summary ofdata reflective of pulp and paper industry progress in reducing the TCDD/TCDF content of effluents,pulps, and wastewater treatment sludges, NY: NCASI, June.
NTP (National Toxicology Program). 1982. Bioassay of 2,3,7,8-Tetrachlorodibenzo-p-dioxin for PossibleCarcinogenicity (Gavage Study), Technical Report Series Number 102. NTP: Research Triangle Park,NC.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 39
OTA. (U.S. Congress, Office of Technology Assessment). 1989. Technologies for Reducing Dioxin inthe Manufacture of Bleached Wood Pulp. OTA-BP-0-54.
OTA (U.S. Congress, Office of Technology Assessment). 1991. Dioxin Treatment Technologies. OTA-BP-0-93. GPO: Wash., DC.
Powell, M. 1996. The 1991 Lead/Copper Drinking Water Rule & the 1995 Decision Not to Revise theArsenic Drinking Water Rule: Two Case Studies in EPA’s Use of Science. RFF Discussion Paper 97-05.Resources for the Future, Wash., DC.
Roberts, L. 1991. “Dioxin Risks Revisited,” Science, Vol. 251, pp. 624-626.
Shabecoff, P. 1987. “Traces of Dioxin Found in Range of Paper Products,” New York Times, Sept. 24, p.A1.
Skrzycki, C. 1997. “Pulp Friction: The EPA’s Tussle Over Paper Pollutant Rules,” The WashingtonPost, April 11, pp. G1-G2.
Smith, B. 1992. The Advisers: Scientists in the Policy Process. Brookings Inst.: Wash., DC.
Thompson, K.M. and Graham, J.D. 1997. “Producing Paper without Dioxin Pollution,” The Greening ofIndustry: A Risk Management Approach, pp. 203-268. Harvard Center for Risk Analysis: Cambridge,MA.
Viscusi, K. 1996. “Alternative Institutional Responses to Asbestos,” Journal of Risk and Uncertainty,Vol. 12, pp. 147-170.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 40
LIST OF ABBREVIATIONS
AET Alliance for Environmental TechnologyAF&PA American Forest and Paper AssociationAOX adsorbable organic halide indicatorAPA American Paper InstituteAWQC ambient water quality criteriaBAT best available technologyBCF bioconcentration factorBCT best conventional (pollution control) technologyBMP best management practiceCDC Centers for Disease Control and PreventionCPSC Consumer Product Safety CommissionCWA Clean Water ActEDF Environmental Defense FundEPA Environmental Protection AgencyFDA Food and Drug AdministrationFOIA Freedom of Information ActFWPCA Federal Water Pollution Control ActIARC International Agency for Research on CancerIJC International Joint CommissionIRIS Integrated Risk Information SystemNCASI National Council of the Paper Industry for Air Stream ImprovementNRDC Natural Resources Defense CouncilNWF National Wildlife FederationOD oxygen delignificationOMB Office of Management and BudgetOTA Office of Technology AssessmentSAB Science Advisory Board, EPATCDD tetrachlorodibenzo-para-dioxinTCDF tetrachlorodibenzofuranTEF toxic equivalent factorTSCA Toxic Substances Control Act of 1976USDA United States Department of Agriculture
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 41
B. Lead in Soil at Superfund Mining Sites
1. Background
This case study deals with an environmental hazard, lead in soil, for which EPAhas no universally-applicable standard that regional and state officials can take from a“look-up table” and “plug in” to determine the “Preliminary Remediation Goal” (PRG) forcontaminated site remediation. Establishing the PRG is the first step that Superfundremediation project managers take in the remedy selection process to begin to address thequestion of “how clean is clean?” Although technical feasibility, cost, the permanence ofthe treatment, applicable or relevant and appropriate requirements (ARARs), and otherfactors are considered by remediation project managers in negotiating the final remedyselection with the PRPs, community, and other agencies, the PRG can serve as animportant benchmark in the negotiations. Because there is no discrete number indicating“safe” levels of lead in soil, this case study does not describe the typical process by whichEPA uses scientific information in site-specific decisionmaking. Instead, it provides someinsight into the challenges and opportunities facing contaminated site programs if theywere to more broadly and consistently develop and use site-specific scientific information.
The Comprehensive Environmental Response, Compensation, and Liability Act,(CERCLA) of 1980 created the EPA Superfund program. CERCLA had two primaryobjectives: to identify and clean up sites contaminated with hazardous substancesthroughout the U.S. and to assign the costs of cleanup directly to those parties--calledresponsible parties--who had something to do with the sites (Probst et al. 1995).Although Superfund is not solely limited to closed or abandoned sites, the program hasprimarily focused on these types of sites. The number of old mineral extraction, milling,and processing sites within the U.S. is unknown, but large, with estimates ranging between100,000 and 400,000 sites. As of 1990, the Superfund National Priorities List (NPL)contained about 1,200 sites, of which sixty were mining sites.47 An additional 220 miningsites are listed by the Comprehensive Environmental Response Compensation and LiabilityInformation System (CERCLIS) (Tilton 1994).48 Lead in soil is a contaminant of concernat over 400 Superfund sites (Ryan and Zhang 1996).
The Resource Conservation and Recovery Act (RCRA) of 1976 deals broadly withhazardous waste produced at operating industrial sites. Mineral extraction and ore millinggenerate two billion tons of solid waste a year, representing nearly 40% of the country’stotal solid waste, and dwarfing the 270 million tons of hazardous waste generated outside
47 As used here, the generic term “mining sites” refers to locations where various stages in the miningprocess occurs. It includes, but is not limited to, sites where mineral-bearing rock called ore is extractedfrom underground or surface mines. During milling (or beneficiation), ore is processed (or upgraded),usually at or near the site of extraction, into concentrates which are further processed in a smelter orrefinery. Final processing sometimes occurs at a considerable distance from the site of extraction.48 CERCLIS contains information on over 30,000 potentially hazardous sites (Tilton 1994). According toa report by the Society of Environmental Geochemistry and Health, as of 1993, soil lead had been found at922 of the 1,300 sites on the NPL (Environmental Health Perspectives, 11/94, pp. 912-913).
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 42
the mineral sector. Responding to concerns of the mining industry, however, Congresspassed the “Bevill Amendment” in 1980 (RCRA Sec. 3001(b)(3)(A)(i-iii)) to exempt“high-volume, low hazard” solid wastes from RCRA Subtitle C hazardous waste federalregulation. Currently, wastes generated from extraction and milling are classified as non-hazardous wastes and subject to state regulation under RCRA Subtitle D, while many (butnot all) mineral processing wastes fall under Subtitle C (Tilton 1994). While much of theongoing discussion concerning lead in soil regulation occurs in the context of theSuperfund program, the potential liability for future RCRA “Corrective Action” motivatesthe owners of active industrial operations to engage in the debate.
Many old mining sites are relatively small and isolated, but the contamination atsome mining and processing sites (including areas in the “fallout zone” downwind ofsmelters) is extensive. The sheer magnitude of many mining sites, relative to mostcontaminated sites, presents EPA with an extraordinary risk management challenge. EPAhas identified approximately 50 “Large Area Lead Sites.” The Jasper County, MOSuperfund Site, for example, includes 10 different areas covering a total of 7,000 acres.49
At a Superfund site located in Midvale, Utah, the old tailings pond alone covers 260 acresand contains 14 million tons of material in uncovered piles.50 In some cases, these largesites overlap residential areas. At the California Gulch Superfund Site in Leadville, CO,mountainous tailings piles lie in close proximity to residences in the small, remote mining
49 Jasper County is an historic lead and zinc mining, milling, smelting area located in southwest Missouri.It is located in EPA Region 7, headquartered in Kansas City, KS. Mining activities in Jasper Countybegan around 1850 and continued through the 1950’s, and millions of tons of mining, milling, andsmelting waste materials are dispersed throughout the county, including residential areas. The site wasadded to the Superfund National Priority List in 1990 (Baysinger-Daniel 1995).50 Tailings are the fine waste particles that are produced during ore milling and typically suspended inwater. Under current practices, tailings from surface mines are deposited in a tailings (or settling) pond,while those from underground mines are deposited in the mine itself. Historically, tailings may havesimply been piled in waste yards or placed in a settling pond. At abandoned sites, however, uncoveredtailings piles often remain after settling ponds have dried up. The Sharon Steel site is located in Midvale, Utah, a small town twelve miles south of Salt Lake City.Around 1906, the US Smelting, Refining, and Mining Company began processing ores at Midvale, and inthe mid-1920s, the company built a mill to process lead-zinc ores and a lead smelter. During World WarII, Utah was a major lead-producing region. ASARCO (American Smelting and Refining Company) alsooperated a lead smelter in Murray, and the International Smelting and Refining Company, a subsidiary ofthe Anaconda Company, had a lead mill and smelter at Tooele. After the war, lead-zinc miningproduction in Utah declined. In 1958, US Smelting and International Smelting agreed to share facilities.US Smelting closed its Midvale smelter and shipped its concentrates to Tooele for smelting andInternational Smelting closed its mill and shipped its ore to Midvale for milling. In 1971, US Smeltingclosed its Midvale mill and changed its name to UV Industries, and in 1979, a year before passage ofCERCLA, the company went bankrupt. US Smelting sold the Midvale site to Sharon Steel, whichplanned to use the site for commercial purposes. The mill and smelter had been torn down, but thetailings pond and slag pile from the smelter remained. The tailings pond contained 14 million tons ofuncovered piles up to 50 feet deep and covered 260 acres. The fine-grained material blew off-site, and44,000 people lived within 2 miles of the site. Health effects were first suspected in 1982, and the site wasproposed for the NPL in 1984 (Tilton 1994). Site characterization is an important element in assessingwhat specific wastes (e.g., lead carbonate or lead sulfide) and proportions of different wastes (e.g., millingwastes vs. smelter emissions) may be present on a particular site. Due to the complex history of manymining sites, this can be difficult.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 43
town. According to Murray (1995), current or historical mining areas of the inter-mountain west often encompass tens or hundreds of square miles.
Although there are a variety of risk management measures that can be taken tolimit exposure to lead-contaminated soil for extended periods (e.g., capping small siteswith asphalt51), the cleanup standards provisions of the 1986 Superfund Amendments andReauthorization Act (SARA, Sec. 121) explicitly state a congressional preference forpermanent treatment. Currently, the only proven, permanent remedial alternative for lead-contaminated soil is excavation and disposal.52 The cost and extent of the cleanup atLarge Area Lead Sites depends on the level or concentration of lead in soil that isconsidered hazardous (i.e., the lower the level of soil lead that EPA considers a healthconcern, the greater the extent and cost of the cleanup). Given the enormous volume ofmaterial involved, excavation and disposal could cost liable parties hundreds of millions ofdollars for some Large Area Lead Sites.53
Opposition to EPA’s proposed remedies for these sites has not come exclusivelyfrom the Potentially Responsible Parties (PRPs), however. In some cases, localgovernments and citizens have objected. The city of Midvale, for example, objected toEPA’s proposed solution because it would have precluded future commercial use of thesite. Leadville residents voiced concerns about property values and jobs.54 At theSmuggler Mountain site near Aspen, CO,55 local residents challenged the need for EPA’sproposed remedy by questioning the basic scientific model that EPA uses for assessing thehealth risks from soil lead (Tilton 1994). (The agency’s “IEUBK Model for Lead inChildren” is discussed in greater detail below.) SCAM (Superfund Coalition AgainstMismanagement), an umbrella group of citizens representing 45 communities around thecountry, filed suit in 1994 against EPA for issuing de facto soil lead cleanup regulations 51 According to an EPA official, the agency has decided to initiate soil excavation at the Jasper CountySite, with the lead-contaminated soils being deposited in the roadbed of a new highway constructionproject. Due to their areal extent, capping Large Area Lead Sites with asphalt is not regarded as a feasiblealternative.52 In this context, disposal means deposition in a facility designed to permanently contain thecontaminated soil. Federal scientists and academics are researching alternative remediation methodsinvolving the application of phosphate, iron oxide, and sewage sludges intended to permanently rendersoil lead less soluble. See, e.g., Ryan and Zhang (1996). An independent toxicologist comments,however, that while phosphate may temporarily immobilize lead, it can mobilizes arsenic, which iscommonly associated with lead on sites contaminated by extractive mining.53 According to an EPA official, the initial remediation cost estimate for the Sharon Steel-Midvale Sitewas in excess of $200 million. All of the sites involved in EPA’s Phase II bioavailability study (discussedbelow) had projected cleanup costs of more than $100 million. Probst et al. (1995) suggests that for themining industrial sector as a whole, the annual financial burden of Superfund cleanup and transactioncosts, estimated at $220.5 million under the current law, is likely to be quite large in relation to theindustry’s profitability.54 According to Murray (1995), Leadville homeowners have suffered a substantial drop in property values.An EPA official estimates that 80% of the employment in Leadville remains mining-related and suggeststhat residents “were parroting whatever the mining company said.”55 The Smuggler Mountain Superfund Site covers approximately 75 acres and is located about 1 milenortheast of Aspen, CO, in Pitkin County. A mobile home park and two condominium developments arelocated a few hundred yards from the site (Lagoy et al. 1989).
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 44
for CERCLA and RCRA sites as OSWER guidance without seeking public input(Environment Reporter, 12/16/94, p. 1608).56
A principal basis for the OSWER soil lead cleanup guidance is the IEUBK model.The agency’s default assumption imbedded in its IEUBK model about how much lead insoil ingested by children would be absorbed into their bloodstream is commonly regardedas driving the cost of the proposed cleanups and has been hotly contested in these cases.In general, the higher the rate of lead absorption, the lower the soil lead PreliminaryRemediation Goal indicated by the IEUBK model. As a result of the standoff over theassumed rate of lead absorption, any cleanup decision regarding the Large Area Lead Sitesthroughout the country was deferred for several years (Tilton 1994) while EPA Region 8(centered in Denver, CO) directed a study on the lead absorbed by juvenile swine afterdigesting soil samples from 8 different sites. Based on previous laboratory studies usingrodents, the presumption by many was that EPA’s model generally overstated the amountof lead that would be absorbed from soil on mining sites, and that it especiallyoverestimated it for mining sites without significant historical milling or smelting activity.The juvenile swine experiment results, however, indicated considerable variability amongthe sites tested, with some higher, some lower, and some about the same as the agency’sdefault assumption, according to an EPA scientist. For at least one site, lead absorptionfrom soils downwind of the smelter was unexpectedly lower than that from soilscontaminated with milling wastes. Consequently, it appears that EPA cannot generalizeacross sites where similar mining activities occurred or even draw any general distinctionsbetween different types of mining sites.
According to an EPA official, the total cost (excluding EPA staff time) of the leadabsorption studies directed by EPA on 8 Large Area Lead Sites was $1.4 million, a smallsum compared to the potential liability associated with the sites. However, soils frommore than 40 sites currently identified as Large Area Lead Sites remain untested, andmany active or old mining sites could become subject to future regulation. Furthermore,once the site-specific lead absorption rate is no longer contested, a new variable in thecomplex risk equation (e.g., local children’s blood lead levels or activity patterns) canreplace it as the disputed factor that drives remediation costs.57
56 CERCLA does not require EPA to conduct national rulemaking. As a result, EPA relies onadministrative guidance to implement Superfund. However, the 1992 Residential Lead-Based PaintHazard Reduction Act (Title X of the Housing and Community Development Act) required EPA todevelop standards defining hazardous levels of lead in lead-based paint, household dust, and soil (addingTitle IV to the Toxic Substances Control Act) by 1994. In July 1994, EPA issued guidance for CERCLAand RCRA (OSWER Directive # 9355.4-12), as well as interim guidance under TSCA. The agency hasnot yet promulgated a formal standard for lead in soil under TSCA. According to press reports, EPAissued TSCA guidance in lieu of a formal rule because existing scientific data could not justify thestandards the agency was considering and suitable data may not be available for five years (Inside EPA,12/16/94, p. 14).57 For example, in the absence of mining activity, background levels of children’s blood lead would tendto be higher than normal in areas with naturally high lead content in the rock and soils. Therefore, oncethe site-specific lead absorption rate is determined, the debate could shift to ascertaining the relativecontribution of the contaminated site to children’s measured blood lead levels. Alternatively, in aprobabilistic analysis of the IEUBK model funded by the Asarco Corporation, Lee et al. (1994) concluded
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 45
Many Superfund critics have characterized the program’s approach to riskassessment as following a “cookbook” and have criticized the program for ignoring site-specific risk assessment information. Under Superfund, EPA, state agencies, orenvironmental consultants often conduct site-specific exposure assessments of varyingscope, depth, and complexity, but the program typically adopts the substance-specificcancer risk values (slope factors), Maximum Contaminant Levels (MCLs), and ReferenceDoses (RfDs) supplied by EPA’s Integrated Risk Information System (IRIS) in awholesale fashion. Superfund generally uses slope factors to assess whether a site’scancer risks lie in the 10-4 - 10-6 “action” range. In accordance with the ARARs provisionof the 1986 SARA, MCLs often determine groundwater remedial objectives. RfD’s areused to evaluate potential non-cancer health effects from soil ingestion.
However, EPA’s current goal for reducing lead health risks is to ensure thatchildren’s blood lead (PbB) levels do not exceed 10 µg/dL (micrograms per deciliter).(Over the past decade the PbB level commonly associated with impairment has decreasedfrom 25 µg/dL to 10 µg/dL). There is no numerical MCL for lead in drinking water, noris there a single numerical RfD for ingested lead.58 Consequently, the Superfund programhas not had the option of following its standard operating procedures for evaluating risksand determining Preliminary Remediation Goals for lead-contaminated sites.59 ThisSuperfund case study, therefore, illuminates the challenges and opportunities posed bydeveloping and using rigorous site-specific scientific information. Table B-1 provides asummary background on Large Area Lead Sites.
that the driving variables include bioavailability of lead and dietary intake of lead. Therefore, once thesite-specific lead absorption rate is determined, the debate could shift to estimating the actual dietaryintake of lead.58 See Powell (1996) for discussion of the development of EPA’s lead policy and regarding EPA’s unusualRfD for arsenic. EPA Region 8 is also directing juvenile swine studies to determine the bioavailability ofarsenic-contaminated soils.59 In principle, Superfund PRGs are developed without consideration of technical feasibility, cost, andpublic acceptance.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 46
Table B-1. Background on Large Area Lead Sites.
1976 RCRA enacted.1980 CERCLA enacted.
Congress passes Bevill amendment to RCRA.1985 EPA concludes that some wastes associated with mineral processing meet the hazardous
criteria for regulation under RCRA Subtitle C, but that the high volumes of waste fromextraction and milling do not.US Centers for Disease Control (CDC) sets screening level for children’s PbB at 25µg/dL,suggests lead in soil or dust appears to be responsible for blood lead levels in childrenincreasing above background levels when the concentration in the soil or dust exceeds 500-1000 ppm.
1986 SARA requires the Dept. of Health and Human Services’ Agency for Toxic Substances andDisease Registry (ATSDR) to prepare a study of lead poisoning in children. After EPA failsto determine which mineral processing wastes would come under the jurisdiction of RCRASubtitle C, the Environmental Defense Fund and the Hazardous Waste Treatment Councilsue the agency.
1988 ATSDR suggests a potential risk of developmental toxicity from lead exposure at PbB levelsof 10-15 µg/dL or lower; identifies paint and contaminated soil as the principal sources oflead for children most at risk (ATSDR 1988).DC Circuit Court of Appeals orders EPA to use the high-volume, low-hazard criteria tonarrow the scope of the RCRA Bevill Amendment exemption for mining wastes.
1989 EPA Office of Solid Waste and Emergency Response (OSWER) guidance recommends a soillead cleanup level of 500-1000 ppm at residential Superfund sites.EPA issues final rules during 1989 and 1990 making most mineral processing wastes subjectto RCRA Subtitle C; however, slag from lead and zinc processing is among 20 mineralprocessing wastes which remain exempt from federal regulation. Extraction and millingwastes remain classified as non-hazardous wastes and thus fall under RCRA Subtitle D.
1990 US Department of Justice files a Superfund suit against Sharon Steel, UV IndustriesLiquidating Trust, and Atlantic Richfield Company (ARCO) regarding a closed leadsmelting facility in Midvale, Utah. ARCO claims that soil lead on the Midvale site poses nohazard. EPA Region 8 directs study rejecting ARCO’s claim. The companies eventuallyagree to pay the government $63 million.
1990 EPA Clean Air Science Advisory Committee (CASAC) recommends a maximum safe PbBlevel for children of 10 µg/dL.OSWER issues RCRA program guidance on soil lead cleanup describing three alternativemethods for setting cleanup levels: 1) use preliminary results of the IEUBK model, 2) use500-1000 ppm range, or 3) use “background” levels at the facility.
1991 EPA formalizes its multi-media Strategy for Reducing Lead Exposures, adopting the10µg/dL PbB level goal.EPA Science Advisory Board (SAB) weakly endorses IEUBK model for developing soil leadcleanup levels at CERCLA and RCRA sites.
1992 Residential Lead-Based Paint Hazard Reduction Act requires EPA to develop standardsdefining hazardous levels of lead in lead-based paint, household dust, and soil.OSWER circulates draft guidance setting 500 ppm lead in soil as the PRGs for Superfundremediations and media cleanup standards (MCSs) for RCRA corrective actions.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 47
Table B-1. Background on Large Area Lead Sites (cont’d).
1994 Urban Soil Lead Abatement Demonstration Project (the 5-year, $15 million “3 Cities Study”)finds a link between soil-lead levels and blood-lead levels in children. However, the studysuggests that soil abatement alone (without lead-based paint stabilization and/or householddust abatement) will have little or no effect on reducing exposure to lead unless there is asubstantial amount of lead in soil (as in Boston test sites, where soil lead levels averaged2,400 ppm) and unless soil lead is the primary source of lead in house dust (Science,10/15/93, p. 323; EPA (1995a)).OSWER issues “revised interim guidance” under CERCLA and RCRA for residential soillead setting a “screening level” of 400 ppm. The screening level is based on the IEUBKmodel using national average inputs and on a goal of limiting exposure to soil lead levelssuch that a child would have an estimated risk of no more than 5% of exceeding the 10µg/dL PbB level. The screening level may be selected as the soil cleanup goal (PRG/MCS),or a site-specific assessment using the IEUBK model can be used to develop soil cleanupgoals. The guidance recommends remedial action when the goal is exceeded, but adds thatsoil excavation may not be necessary. Research to determine the bioavailability of soil leadis encouraged for mining sites without significant past milling/smelting activity.Citizens’ group sues EPA for issuing de facto regulation as guidance without publiccomment.
1995 House Commerce Committee Chair Thomas Bliley (R-VA) expresses concern about EPA’ssoil lead cleanup guidance and objects to the agency’s reliance on the IEUBK model tojustify the lead policy (Environmental Executive Report, 10/30/95).EPA proposes to tighten controls of many mineral processing wastes previously exempt fromRCRA Subtitle C.
1996 Results of juvenile swine studies directed by EPA Region 8 indicate considerable variabilityin soil lead bioavailability among Large Area Lead Sites.
2. Scientific Issues
NRC (1993) reports that the health effects of lead at approximately 10 µg/dL inblood include:
• impaired cognitive function and behavior in young children• increases in blood pressure in adults, including pregnant women• impaired fetal development• impaired calcium function and homeostasis60 in sensitive populations
NRC (1993) also concludes that somewhat higher PbB concentrations are associated withimpaired biosynthesis of heme (a substance required for blood formation, oxygentransport, and energy metabolism) and cautions that some cognitive and behavioral effectsmay be irreversible.
60 This refers to maintaining calcium at appropriate levels in the body.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 48
The primary commercial use for lead is making batteries.61 In 1990, the U.S.produced (excluding recycling) approximately 500 thousand tons of lead (Young 1992).Environmental releases of lead from the minerals industry results not only from mininglead and lead-zinc ores, but more generally from mining and smelting of non-ferrous(noniron) metals. Copper mining, for example, can release lead into the environment.Like all heavy metals, lead accumulates in the environment because it does not degrade.In soils, lead tends to accumulate in surface organic matter (Kabata-Pendias and Pendias1984). In contrast to EPA’s soil lead screening level of 400 ppm, the lead content ofagricultural soils ranges from 1 to 135 ppm with a typical value of 10 ppm (Ryan andZhang 1996). As a result of industrial and automotive emissions and exterior lead paint,inner-city neighborhoods in many of our major cities have elevated accumulations of leadin soil.62 Because lead is reported to be least mobile in soils among the heavy metals (Kabata-Pendias and Pendias 1984), the leaching of lead into groundwater is generally of less concernthan is direct ingestion of soil lead, particularly since children frequently engage in hand-to-mouth activity.63 According to an EPA scientist, lead and arsenic are generally theprincipal contaminants of concern to human health at mining sites.
During the mid-1980s, EPA began developing a risk assessment model that strivesto take into account simultaneously the various pathways of lead exposure--inhaled autoand industrial emissions and ingested soil, dust, food, and drinking water.64 The model,referred to as the Integrated Exposure Uptake Biokinetic Model for Lead in Children (orIEUBK Model), was developed by the Office of Air and Radiation Office of Air QualityPlanning and Standards (OAR/OAQPS), first in the context of reviewing the NationalAmbient Air Quality Standard for lead. The model was later used for the Lead/Copperdrinking water rulemaking (see Powell (1996)), and is currently managed by the EPAOffice of Research and Development’s National Center for Environmental Assessment(ORD/NCEA). The model predicts a distribution of children’s blood lead levels as afunction of existing blood lead levels, inputs from the various sources, and different lead“bioavailability” rates. Bioavailability refers to the rate and extent of absorption of asubstance. For ingested substances, it is measured by the fraction of the orally
61 Lead is also used for radiation shielding, cable covering, ammunition, chemical reaction equipment, fusiblealloys, type metal, vibration damping in heavy construction, foil, and bearing alloys (Hawley 1981). Leadedgasoline, new lead plumbing, and the pesticide lead arsenate have been phased out.62 Ryan and Zhang (1996) report that many inner-city neighborhoods have mean or median soil leadconcentrations in excess of 1000 ppm, with values as high as 50,000 ppm being reported. However, anOSWER official questions whether these values are overestimates. By comparison, the Urban Soil LeadAbatement Demonstration Project was performed in areas with median soil lead concentrations rangingfrom 237 - 2,396 ppm (EPA 1995a). Suffice to say that levels of lead in soil in many urban areas,particularly in older neighborhoods, are considerably higher than background levels due to accumulatedanthropogenic releases of lead into the urban environment.63 However, lead does become more soluble in acid soils (Kabata-Pendias and Pendias 1984), and sulfides(which can produce sulfuric acid) make up more than a third of many nonferrous mineral ores (Young 1992).64 According to EPA officials, the IEUBK model builds on research conducted in the late 1970s at NewYork University by Harley and Kneip, who developed a pharmacokinetic model for lead (describing theabsorption, movement, storage, and excretion of lead in various compartments of the body) by conductingstudies with juvenile baboons and by analyzing human cadavers. The IEUBK added exposure components(levels of lead in air, food, water, soil, and dust) to the pharmacokinetic model.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 49
administered dose that is absorbed by the body.65 The default assumption for thebioavailability of ingested lead in soil and dust under the IEUBK Model is 30%.66
According to Murray (1995), site-specific risk estimates for contaminated mining areas aresensitive to variability in assumed bioavailability parameter values. Other things beingequal, the higher the rate of lead absorption, the lower the level of soil lead that presents ahealth concern (and, potentially, the greater the extent and cost of cleanup).
Bioavailability of lead is a function of a number of factors, including the solubilityof the chemical form of lead ingested, the soil particle size and surface area-to-weightratio, the pH of the digestive tract, and nutritional status. The physical and chemicalcharacteristics of soil lead vary from site to site. Based largely on the argument that thesoil lead at the Smuggler Mountain Superfund Site “is almost certainly less bioavailable”than the “lead present in soils in the vicinity of smelters and in urban areas where most ofthe studies relating the concentration in soil to body burdens have been conducted,” agroup of environmental consultants estimated that soil lead levels of 1,000 ppm would be“safe” (Lagoy et al. 1989).67 Findings reported by Steele et al. (1990) and Hemphill et al.(1991) suggested that lower-than-default soil lead bioavailability values may beappropriate generally for mining sites, particularly where the predominant source of leadcontamination was not late-stage processing activities such as smelting and milling thatproduce very fine particle sizes.
Lead absorption also varies with age. It is higher for children than for adultsbecause calcium and lead absorption are linked. Children absorb calcium at high ratesbecause their bones are growing, but the human metabolism does not distinguish wellbetween lead and calcium, so lead unfortunately “goes along for the ride.” As a result, themost common laboratory animals, mature rodents, are not the best animal model forassessing lead bioavailability if physiological similarity to children is the primary evaluativecriterion.68 Emphasizing this criterion, EPA Region 8 selected juvenile swine as theappropriate animal model for testing soil lead bioavailability on Large Area Lead Sites.
Swine, however, have some distinct disadvantages as laboratory animals. Notably,they are large (approximately 35 lbs.) and expensive. In contrast to the abundance of
65 The absorption rate can be a function of administered dose. For example, according to an independenttoxicologist, the relationship between blood lead and dose becomes curvilinear at higher doses due toeither attenuated uptake or more efficient excretion.66 By contrast, the IEUBK default bioavailability for the generally more soluble forms of lead present infood and water is 50%.67 Lagoy et al.. (1989) based their determination partially on 1985 CDC guidance which suggested thatincreased blood lead levels were associated with soil lead levels of 500-1000 ppm. Lagoy et al.. (1989)also indicate that various approaches to establishing a “safe” level of lead in the soil yield widely varyingresults ranging from about 100 ppm to 8000 ppm.68 According to a federal scientist, juvenile swine have been found to be better animals to model humanuptake of nutrients and carbohydrates. An EPA scientist notes that rodents mature at 6-10 weeks, whereasswine mature more slowly. Other problems with the rodent model include a low “residence time” for food(ingested substances pass through the animals quickly, leaving little time for lead absorption) and the factthat rats eat their feces, resulting in a confusing “redosing” of the animals with excreted lead.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 50
laboratories certified to conduct tests with small animals (e.g., rodents, rabbits, and fish),there are very few facilities across the country that can provide the necessary qualityassurance to conduct reliable laboratory studies using large animals.69 A soils scientistnotes that rodent studies cost 10 times less, on a per animal basis, than juvenile swinestudies (the cost per soil sample is $5,000 and $50,000, respectively). Consequently, for agiven sum of money, researchers using rodents can conduct a larger, more statisticallypowerful study (i.e., one able to detect smaller differences between treatment levels).70
However, the rodent studies may systematically underestimate soil lead bioavailability inchildren. According to an EPA scientist, Superfund PRPs promoted the rodent model,because “conveniently for them, it showed low absorption, but that’s expected because
To a considerable extent, selection of the most appropriate animal model involvestradeoffs between cost, experimental power and control, fidelity to children’s physiology,and the value of information for decisionmaking. Determination of the “optimal” animalmodel may depend on which evaluative criterion is being used. According to a soilsscientist, when the US Department of Agriculture (USDA) conducts nutritional research,it begins with rodent studies to contain research costs, and then, if deemed necessary,moves up progressively to “higher,” more costly animals, such as swine or primates.Underscoring the subjectivity involved, this source remarks that in the future, using theresults of rodent studies calibrated by the available swine data “is good enough for me.With all of the lead and zinc mine wastes [that need to be tested], it will save millions [ofdollars].”
Currently, researchers are working to develop cheaper in vitro laboratory modelsto study lead bioavailability.71 Most parties seem to agree that in vitro methods areultimately needed to get around the need for animal research altogether. But until acheap, reliable, and valid in vitro model is developed, disputes over selection of the mostappropriate animal model will linger. Disagreement among scientists about thephysiological suitability of the competing animal models has largely subsided, and opinionamong toxicologists has apparently converged on the juvenile swine as an appropriate testanimal for high stakes Superfund decisions. But in the early 1990s, when EPA Region 8initiated the swine studies, the bulk of previous work on the bioavailability of heavy metalsin soils had been conducted with rodents. According to an EPA scientist, “otherinvestigators in the field were challenged professionally by [EPA’s] decision to use a largeanimal model.” Selection of the appropriate animal model was hotly contested not only byscientists, but also by PRPs and their environmental consultants.
69 The juvenile swine studies were conducted at the University of Missouri under contract with EPA.70 An EPA scientist suggests that the precision of the swine studies has been greater than expected.71 Scientists at the University of Colorado Geochemistry Department and the EPA/ORD/ECAO laboratoryin Cincinnati, OH are working to develop a “fake GI [gastrointestinal] tract,” says an EPA scientist. Thebenchtop model consists of a series of chambers with flow-thorough mechanisms, and preliminary resultshave been “pretty consistent” with those from the juvenile swine model.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 51
3. Process Within EPA
Setting the Agenda
According to an OSWER official, the program was first introduced to the IEUBKmodel by Jeff Cohen, who was responsible for developing the model as a member of theOAQPS staff and later directed the Office of Drinking Water’s Lead Task Force duringfinalization of the 1991 Lead/Copper Drinking Water Rule. “Our usual approach,” saysthe OSWER official, “is to use RAGS [the Risk Assessment Guidelines for Superfund],site-specific exposure assessments, and then go to IRIS [the Integrated Risk InformationSystem] for cancer slope factors, MCLs, and RfDs. For lead, those were not available tous in IRIS, and we needed some other way to evaluate the sites.”
EPA Region 8 first applied the IEUBK model in a Superfund context in the late1980s on a site in E. Helena, MT where there was a closed lead smelter.72 According toan EPA scientist, data collected from the site were used to develop agency assumptions inthe soil/dust ingestion component of the IEUBK model. For the purposes of the E.Helena remedy selection, the results of the IEUBK model were not used to generate thefinal numerical soil cleanup level. Instead, “the model was used as a means of technicalnegotiations with the PRP, ASARCO (American Smelting and Refining Company),” saysan EPA official. “EPA was quite successful with it under those circumstances,” in partbecause one of the ASARCO technical consultants had worked academically indevelopment of the model.
Phase I
In 1989, OSWER issued guidance (OSWER Directive #9355.4-02) recommendinga soil lead cleanup level of 500-1000 ppm73 at Superfund sites where the land wascurrently in residential use or where remediation project managers believe futureresidential use is possible.74 Late in the same year, the federal government was preparingfor litigation on the Sharon Steel Superfund Site in Midvale, Utah. The site historicallyhad been a lead smelting operation, and there were also mining wastes present on thelocation. One of the PRPs, the Atlantic Richfield Company (ARCO), argued that the
72 In 1986, the US Centers for Disease Control and the EPA issued a report, East Helena, Montana, ChildLead Study, Summer 1983.73 This range is identical to that in the 1985 CDC guidance to which Lagoy et al.. (1989) referred.74 The role of landuse in Superfund cleanups has been a source of considerable controversy.EPA/OSWER (1989, p. 6-7) directs personnel to “assume future residential land use if it seems possiblebased on the evaluation of the available information.” The presumption of residential land use affects riskassessment exposure assumptions regarding levels of on-site soil ingestion, and it may also affectassumptions regarding the quality of drinking water drawn from groundwater. According to EPA (1994),residential landuse is the fourth most common current landuse at Superfund sites and the second mostcommon expected future landuse. According to EPA (1995b), 15% of Superfund sites have people livingon site, 80% have residence adjacent to them, and about 20% of the time, when a commercial/industrialsite has residences nearby, EPA will assume a future residential land use. See Probst, Wernstedt, andHersh (1996) for a discussion of the many definitions of land use and its role in Superfund cleanups.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 52
particular form of soil lead at the site was not hazardous because it had zerobioavailability. The Department of Justice, which is responsible for litigating Superfundcases in court, approached a toxicologist with EPA Region 8, Chris Weis, to ascertainwhether ARCO could be right. Weis recommended a short-term experiment usingjuvenile swine to test ARCO’s argument. The results of the study (now referred to asPhase I) rejected ARCO’s assertion that the soil lead had zero bioavailability, but alsosuggested it was somewhat lower (20-25%) than EPA’s default assumption under theIEUBK model (30%).75
Unbeknownst to EPA and the Justice Department, ARCO had already conductedsome lead bioavailability tests on soil samples from the Midvale site using rodents, and thePRP felt that it had a rock-ribbed scientific argument for minimizing its liability.According to an EPA scientist, “ARCO was furious that we criticized the studies they had
76 The PRP may have been surprised by EPA’s challengebecause although EPA Regional Offices are generally staffed with a complement ofenvironmental engineers and hydrogeologists, they typically contain relatively few expertsin the health aspects of environmental risk assessment, such as Weis.77 An EPA officialobserves, “They [ARCO] were for the first time faced with a trained toxicologist whocould challenge their evidence. Now they had to contend not only with lawyers andengineers, but also with trained scientists.” The PRPs “brought their checkbook to court”and settled on the first day of the trial for $63 million. Prior to the swine study, theestimated cleanup cost was more than $200 million, according to an EPA official.78
The resolution of the Sharon Steel case did not, however, clear up the disputeabout EPA’s use of the IEUBK model for setting numerical soil lead cleanup goals, eitherfor the national Superfund program or for any particular contaminated site. Instead, theterms of debate shifted to attempts to distinguish among types of Large Area Lead Siteswhere soil lead was expected to be more or less bioavailable (thus obviating the need forcostly and time-consuming site-specific applied research) and to arguments over theappropriate laboratory animal for use in testing soils for lead bioavailability (i.e., rats vs.pigs).
75 OSWER Headquarters and ORD/ECAO each provided $100,000 for Phase I.76 According to an independent toxicologist, external scientists such as Herbert Needleman of theUniversity of Pittsburgh, John Drexler of the University of Colorado, and Paul Mushak of PB Associateswere also prepared to testify as expert witnesses on behalf of the government in the Sharon Steel case.Mushak had submitted a critique of ARCO’s arguments that laboratory solubility tests approximated thebehavior of lead tailings in humans. Drexler had performed laboratory analyses of Midvale soil samplesindicating that lead was present in very fine particles, and therefore more likely to be bioavailable, andalso consisted of toxic chemical forms. See Powell (1996) for further discussion of this case andNeedleman’s role.77 Weis has a Ph.D. in environmental toxicology, is a board certified toxicologist, and conducted post-doctoral research in physiology and biophysics at the University of Virginia Medical School prior tojoining EPA Region 8. According to an OSWER official, Regions 1 and 8 have a comparatively largenumber of health scientists.78 Tilton (1994) estimates that the transaction costs of doing the technical work, negotiating and litigatingthe case accounted for between a fifth and a third of the total funds devoted to the Midvale site by thePRPs and EPA.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 53
Phase II
Although OSWER had issued broad soil lead guidance for Superfund and RCRAin 1989-90, pressure was mounting for EPA to provide more definitive guidance for lead-contaminated soil cleanups. In 1992, the Residential Lead-Based Paint Hazard ReductionAct required that EPA develop national standards for hazardous levels of lead in soil.Many EPA remediation project managers and PRPs were hopeful that the Sharon Steelsite test results indicating lower-than-IEUBK-default soil lead bioavailability could beapplied to the remainder of the Large Area Lead Sites. Others, including scientists whohad conducted previous heavy-metal bioavailability research and some PRPs, continued todefend the application of rodent studies. However, EPA Region 8 officials cautioned thata decision to apply the Sharon Steel site test results (Phase I) to all mining sites could notbe supported technically due to the variability in site conditions.79 According to Murray(1995), the PRPs for the Leadville site submitted the results of rodent studies with lower-than-default bioavailability values. Given the results of Phase I, Region 8 was reluctant todepart from the default on the basis of rodent studies. Subsequently, a second tranche(Phase II) of juvenile swine studies was conducted testing lead bioavailability in soilsamples from 7 different Superfund sites across the country.80
For some sites, the results indicated bioavailability rates substantially lower thanthe IEUBK default of 30%. For example, the bioavailability was 19% at the BinghamCreek Superfund Site near Salt Lake City, Utah (Murray 1995). The studies have not,however, consistently yielded bioavailability rates lower than 30%. According to EPAofficials, the Leadville site results were not substantially lower than the default. For theJasper County site, the bioavailability of lead in soils contaminated by smelting emissionswas in the 28-31% range. Surprisingly, the bioavailability of soil lead from mining wasteswas higher--in the 40-50% range. This finding was contrary to the conventional wisdomthat smelter fallout would produce the most highly bioavailable lead form present onmining sites.
The Role of External Scientists in the Process
After holding a 1988 conference on lead in soil, the Society for EnvironmentalGeochemistry and Health (SEGH) formed a “Lead in Soil” Task Force to examine therelationship between blood lead and environmental lead and to develop guidelines forassessing and managing the health risks associated with lead in soil and dust. (For adiscussion of the role of SEGH in the arsenic in drinking water case, see Powell (1996).)In 1990, the SEGH, with support from EPA and other sponsors, convened a scientificmeeting in Chapel Hill, NC on bioavailability and dietary intake of lead. The SEGH task
79 Phase I was also designed simply to quickly test the hypothesis that the soil lead bioavailability at theMidvale site was zero. A larger, more elaborate test was required to estimate the actual bioavailabilitywith much precision.80 Phase II studies were supported by regional Superfund program budgets (with cost recovery provisionsapplicable).
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 54
force report (summarized in Wixson and Davies 1994) questioned the use of a single valuefor lead in soil for use in all cleanup situations. An EPA scientist says, “I got the feelingthat it [SEGH] was somewhat dominated by the regulated industry. But to their credit,nobody else was providing a forum for these discussions.” The major impact that SEGHhad was in focusing EPA’s attention on the need to consider differences among sites insoil lead bioavailability.81
In 1991, a year after OSWER issued RCRA program guidance permitting the useof the IEUBK model for setting soil lead cleanup levels, the EPA Science Advisory Board(SAB) concluded that the IEUBK model represented an improved methodology forassessing total lead exposure and for developing soil lead cleanup levels at CERCLA andRCRA sites. However, the SAB raised concerns about incorrect application of the modeland selection of inappropriate input values for default and site-specific applications. Inresponse, OSWER developed a guidance manual for the IEUBK model.
4. Science in the Remedy Selection
The bioavailability of lead in soil isdirectly related to the potential extent and costof Large Area Lead Site cleanups. Other factorsbeing equal, the higher the bioavailability, thelower the soil lead level indicated by the IEUBKmodel for the preliminary remediation goal(PRG). As indicated above, a suite of factors isconsidered in the remedy selection process, buta lower soil lead PRG suggests that a greatervolume of dirt would have to be excavated anddisposed of to permanently meet the health-based cleanup goal using currently provenremedial technologies. The juvenile swine studies estimated soil lead bioavailability toprovide a site-specific refinement of the IEUBK model. Perhaps not surprisingly, theresults indicated considerable variability among the sites tested, with some higher, somelower, and some about the same as the agency’s default assumption. In terms of the finalremedy selection, however, it appears that none of the results will increase the liability ofLarge Area Lead Site PRPs because EPA deems the cost of removing the contaminatedsoil to be excessive.
According to an EPA official, although agency scientists had expected the swinestudy for the Midvale site to indicate even lower soil lead bioavailability, the differencebetween the study results (20-25%) and the default assumption (30%) “was significant in
81 SEGH task force member Rufus Chaney, a USDA researcher, was a strong proponent of using rodentsfor testing lead bioavailability. Chaney’s bioavailability studies with rodents were supported by Dupont.According to an EPA official, DuPont’s interest stems from the company’s interest in bioavailability ofresiduals and alternative remediation technologies. EPA was represented on the task force by RichardCothern (see Powell (1996) for discussion of Cothern’s role in the scientific assessment of lead andarsenic in drinking water).
It appears that regardless of whethersite-specific experiments indicate asoil lead bioavailability higher orlower than the agency’s defaultassumption, the results will notincrease the liability of Large AreaLead Site PRPs because EPA deemsthe cost of removing thecontaminated soil to be excessive.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 55
terms of site liability,” causing EPA “to raise the proposed action level” (i.e., the soil leadconcentration of concern) for the site. For the Bingham Creek Site near Salt Lake City,the experimentally estimated soil lead bioavailability was 19%. According to Murray(1995), the reduced bioavailability value halved the estimated cost of site cleanup from $8million to $4 million. According to an EPA official, it is questionable whether the slightlyreduced soil lead bioavailability estimate for the Leadville site was big enough to make animpact on the risk management decision.
For the Jasper County site, as a result of the unexpectedly higher bioavailability ofmining wastes relative to smelter fallout, the estimated geographic distribution of effectiveexposures to lead via soil was markedly different than initially suspected, according to anofficial with the Missouri Department of Natural Resources. Consequently, while theswine study did not affect the soil cleanup level, it altered the geographic emphasis ofcleanup activities. According to an EPA official, the agency has decided to initiate soilexcavation at the site while studying the feasibility of treating the soil with phosphate toreduce the soil lead solubility. According to a Missouri state health official, public demandfor additional swine studies to determine the bioavailability of soil lead after phosphatetreatment has helped overcome some bureaucratic reluctance to allocating the necessaryfunds. Although the juvenile swine study did not dramatically reduce the scale of remedialactivity, the EPA official considers the study “money well spent. It settled a lot ofdisputes and saved a lot of arguing regarding what the real number was.” Thus, animportant contribution of the swine studies to the Jasper County site has been theircapacity as a dispute resolution tool (perhaps lowering transaction costs). The resultshave been accepted as a means for determining the remedial action plan and for assessingthe performance of an alternative treatment.
Despite the fact that tests have estimated soil lead bioavailability to be higher thanthe IEUBK default on some sites, the information has been either beneficial or essentiallyneutral to Large Area Lead Site PRPs, apparently because EPA feels that the cost ofremoving the contaminated soil is excessive. Regarding the sites where bioavailability maybe higher than the default, an EPA official remarks, “It could potentially lead to higherliability, but I doubt it. All the discussion is about lowering the default.” Regarding LargeArea Superfund Sites generally, an EPA official concludes, “we will be managing exposure[to environmental hazards] in perpetuity” due to the prohibitive costs of permanentlyremoving the risk. As a result of EPA’s implicit policy, the PRPs can take a chance ondramatically reducing their costs by exposing themselves to a relatively modest increase insite liability (EPA can recover the swine study costs), secure in the knowledge that theyare not exposing themselves to greater financial risk, even if the science “goes the wrong
5. Concluding Observations
In terms of a fate and transport analogy, this case study again illustrates themobilization of science (in particular, the IEUBK model) accumulated in differentcompartments of EPA (the air and drinking water programs) and its assimilation by
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 56
another compartment (the CERCLA and RCRA programs). While the soil leadbioavailability studies for Large Area Lead Sites were being done, most of the attention atEPA headquarters at the time was on the high-profile Urban Soil Lead AbatementDemonstration Project (EPA 1995a), according to an OSWER official. Thus, EPARegion 8 also was able to exploit the institutional learning that was occurring in thecontext of the Urban Soil Lead Abatement study for the purposes of the lead mining siteswithout the attendant supervision and scrutiny. As in other case studies conducted as partof this project, individual EPA staff (Cohen and Weis) who could bridge media-boundprogram areas and interface with scientists played key roles in ensuring that the sciencegenerated by the agency was put to use in regulatory decisions. Cohen integrated whatwas known about the effects of lead exposure from various sources (air, drinking water,and soil) that were artificially disaggregated by EPA’s organizational structure into threeseparate programs (the criteria air pollutants, drinking water, and Superfund/RCRA), eachwith their own parochial concerns. As a trained health scientist on staff at an EPAregional office, Weis was able to design an experiment to test the scientific informationgenerated by the regulated community and deposited in the office. These dynamics areillustrated in Figure B-1.
Figure B-1. Fate and Transport Dynamics for Science in the Lead in Soil at Mining SitesDecisions.
Due to the limited number of experts in the health aspects of environmental riskassessment in EPA regional offices, the current management options often may be limitedto either absorbing or rejecting such information produced by Potentially ResponsibleParties. Relatively few health scientists may be needed if EPA regional offices (and stateagencies with delegated responsibility) are expected to uniformly apply the substance-specific toxicological values provided by IRIS and other central databases. If, on theother hand, environmental agencies are expected to more consistently use site-specificscientific information in the contaminated site remedy selection process, they face thechallenge of having adequate scientific capacity available (through an appropriate mix ofstaff and consultants) to manage the production of site-specific data and to criticallyevaluate the information produced by the regulated community.
EPAAir and
DrinkingWater
Programs
EPA Region 8
IEUBK Model
Large AreaLead Site
PRPs
Mature Rodent Studies
Juvenile Swine Studies
Default: SoilLead 30%
Bioavailable
Soil Lead at Mining SitesLess than 30% Bioavailable
Soil LeadBioavailability at
Mining Sites Varies ona Site-by-Site Basis
Large Area Lead SitePRPs’ Liability Does
Not Increase
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 57
REFERENCES
ATSDR (Agency for Toxic Substances and Disease Registry). 1988. Nature and Extent of ChildhoodLead Poisoning in Children in the US: A Report to Congress. US Department of Health and HumanServices. July.
Baysinger-Daniel, C. 1995. Cooperation and Coordination among Several Agencies and Private Groupsat the Jasper County Superfund Site. Poster Presentation. State and Tribal Forum on Risk-BasedDecision Making, St. Louis, MO, November 12-15.
EPA/OSWER (U.S. Environmental Protection Agency/Office of Solid Waste and Emergency Response).1989. Risk Assessment Guidance for Superfund. Volume I. Human Health Evaluation Manual.EPA/540/1-89/002.
Hawley, G. 1981. The Condensed Chemical Dictionary, 10th ed. Van Nostrand Reinhold Company: NY.
Hemphill, C., M. Ruby, B. Beck, A. Davis, and P. Bergstrom. 1991. “The Bioavailability of lead inMining Wastes: Physical/Chemical Considerations,” Chemical Speciation and Bioavailability, vol. 3, pp.135-148.
Kabata-Pendias, A. and H. Pendias. 1984. Trace Elements in Soils and Plants. CRC Press, Inc.: Boca Raton,Fl.
Lagoy, P., I. Nisbet, and C. Schulz. 1989. “The Endangerment Assessment for the Smuggler Mountainin D. Paustenbach, ed. The Risk Assessment of
Environmental and Human Health Hazards: A Textbook of Case Studies. John Wiley & Sons: NY, pp.505-525.
Lee, R., W. Wright, and W. Haerer. 1994. Development of a Stochastic Blood Lead Prediction Model.Abstract from 1994 Society for Risk Analysis Annual Meeting.
Murray, B. 1995. Bioavailability of Lead in a Juvenile Swine Model to Support Superfund Mine SiteReclamation. Presentation at the State and Tribal Forum on Risk-Based Decision Making, St. Louis, MO,November 12-15.
NRC (National Research Council). 1993. Measuring Lead Exposure in Infants, Children, and OtherSensitive Populations. National Academy Press: Wash., DC.
Powell, M. 1996. The 1991 Lead/Copper Drinking Water Rule and the 1995 Decision Not to Revise theArsenic Drinking Water Rule: Two Case Studies in EPA’s Use of Science. RFF Discussion Paper 97-05.Resources for the Future: Wash., DC.
Probst, K., D. Fullerton, R. Litan, and P. Portney. 1995. Footing the Bill for Superfund Cleanups: WhoPays and How? The Brookings Institution and Resources for the Future: Wash., DC.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 58
Probst, K., K. Wernstedt, and R. Hersh. 1996. The Role of Landuse in Superfund Cleanups.Forthcoming report from Resources for the Future: Wash., DC.
Ryan, A. and P. Zhang. 1996. Soil Lead Remediation: Is Removal the Only Option? Research paper byUS EPA Risk Reduction Engineering Laboratory, Cincinnati, OH,http://128.6.70.23/html_docs/rrel/ryan.html.
Steele, M., B. Beck, B. Murphy, and H. Strauss. 1990. “Assessing the Contribution from Lead in MiningRegulatory Toxicology and Pharmacology, Vol. 11, pp. 158-190.
Tilton, J. 1994. “Mining Waste and the Polluter-Pays Principle in the United States,” in R. Eggert, ed.Mining and the Environment. Resources for the Future, Wash., DC, pp. 57-84.
Wixson, B. and B. Davies. 1994. “Guidelines for Lead in Soil: Proposal of the Society forEnvironmental Geochemistry and Health,” Environmental Science and Technology, Vol. 28, No., 1, pp.26A31A.
Young, J. 1992. “Mining the Earth,” State of the World 1992. Worldwatch Institute: Wash., DC, pp.100-118.
Dioxin from Pulp and Paper and Lead in Soil at Mining Sites 59
List of Abbreviations
ARCO Atlantic Richfield CompanyASARCO American Smelting and Refining Company, IncorporatedCDC Centers for Disease Control and PreventionCERCLA Comprehensive Environmental Response, Compensation and
Liability Act of 1980CERCLIS Comprehensive Environmental Response Compensation and
Liability Information SystemECAO Environmental Criteria and Assessment Office, EPAEPA Environmental Protection AgencyIEUBK Integrated Exposure Uptake Biokinetic (model for lead in children)MCL Maximum Contaminant Levelµg/dL micrograms per deciliterNPL National Priority ListOAQPS Office of Air Quality Planning and Standards, EPAOAR Office of Air and Radiation, EPAORD Office of Research and Development, EPAOSWER Office of Solid Waste and Emergency Response, EPAPbB blood leadppm parts per millionPRG preliminary remediation goalPRP potentially responsible partyRCRA Resource Conservation and Recovery Act of 1976RfD reference doseSAB Science Advisory Board, EPASARA Superfund Amendments Reauthorization Act of 1986SEGH Society of Environmental Geochemistry and Health
